JACOBSON et al v. DETZNER
Filing
202
FINAL ORDER FOLLOWING BENCH TRIAL. The Clerk shall enter judgment stating: "This Court hereby DECLARES that the ballot order scheme described in section 101.151(3)(a), Florida Statutes, violates Plaintiffs' rights under the First and Fourteenth Amendments of the United States Constitution. This Court GRANTS Plaintiffs' request for a permanent injunction. Pursuant to the Secretary of State's responsibility for "general supervision and administration o f the elections laws," § 15.13, Fla. Stat. (2019), and her authority to "[o]btain and maintain uniformity in the interpretation and implementation of the elections laws," § 97.012(1), Fla. Stat. (2019), and pursuant to Federa l Rule of Civil Procedure 65(d)(2)(C), neither Defendant Lee, her successors in office, deputies, officers, employees, agents, nor any other person in active participation or concert with Defendant Lee shall enforce, nor permit enforcement of, the ba llot order scheme described in section 101.151(3)(a), Florida Statutes. Defendant Lee and her successors in office, as well as her deputies, officers, employees, agents, and any other person in active participation and concert with Defendant Lee shal l take all practicable measures within the scope of their official authority to ensure compliance with the terms of this Order. From the date of this Order forward, no ballot shall issue which is organized pursuant to the ballot order scheme descri bed in section 101.151(3)(a), Florida Statutes. No supervisor of elections of any Florida county, nor their successors in office, deputies, officers, employees, agents, nor designees, shall issue any ballot which is organized pursuant to the ballot o rder scheme described in section 101.151(3)(a), Florida Statutes." Within fourteen days after the issuance of this Order, Defendant Lee shall provide written guidance to the supervisors of elections of Florida's counties informing them t hat this Court has declared the ballot order scheme described in section 101.151(3)(a), Florida Statutes, unconstitutional. Once Defendant Lee has provided said written guidance to the supervisors of elections, Defendant Lee shall file a notice of compliance in this Court within twenty-one days after issuance of this order. Pursuant to this Court's authority to monitor the status of its injunction, when the State of Florida adopts a new permanent ballot order scheme to replace the ballot order scheme described in section 101.151(3)(a), Florida Statutes, Defendant Lee shall file a notice in this Court informing this Court that the State of Florida has done so, and shall append thereto a copy of the final text of the ballot order scheme adopted. The Clerk shall close the file. Signed by CHIEF JUDGE MARK E WALKER on 11/15/2019. (kjw) Notice of Compliance by Defendant Lee to be filed by 12/6/2019.
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 1 of 74
IN THE UNITED STATES DISTRICT COURT
NORTHERN DISTRICT OF FLORIDA
TALLAHASSEE DIVISION
NANCY CAROLA JACOBSON,
et al.,
Plaintiffs,
v.
CASE NO.: 4:18cv262-MW/CAS
LAUREL M. LEE, et al.,
Defendant/Intervenors.
_______________________________/
FINAL ORDER FOLLOWING BENCH TRIAL
In political circles, it is widely believed that the candidate who is listed first
on the ballot has an advantage in the election—an advantage which can be decisive.1
This advantage supposedly comes from a phenomenon called the primacy effect,
which is the human tendency to choose the first item in a list of options. The portion
of the vote a candidate supposedly gains this way is called the “primacy effect” vote,
the “windfall vote,” or the “donkey vote.”2 See Sarvis v. Judd, 80 F. Supp. 3d 692,
699 (E.D. Va. 2015).
1
It bears noting that Intervenors—a group of Republican organizations—originally sought
to intervene in this case on the basis that Republican candidates and organizations “stand to be
most directly harmed by a change” in Florida’s ballot order scheme, although they now argue it
awards no significant advantage to any candidate or party. ECF No. 23 at 16.
2
Here, “donkey” probably substitutes for a more colorful term. See Richard Leighton,
Don’t be a smartass and confuse donkeys, mules, The Ellsworth American (July 12, 2019)
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 2 of 74
In Florida, the order in which candidates appear on the ballot is determined
by the previous election for governor. The party in control of the Governor’s
Mansion has its candidates listed first in every race, all along the ballot. The party
that came in second has its candidates listed in the second position in each race. So,
if the Governor of Florida is a Democrat, then Democrats will be listed first in every
race on every ballot for the next four years, just because they are also Democrats.
The implication is obvious. Assuming the so-called “donkey vote” exists,
Florida’s ballot order statute ensures one party’s candidates receive that advantage
in every race, all down the ballot, in every election. In practical terms, when the
governor is a Democrat, this means every Democratic candidate has a small but
significant advantage in every election over the Republican candidate, and that they
have this advantage solely because they are Democrats. When a Republican is
governor, Republicans have the advantage, solely because they are Republicans.
The first issue in this case is whether Plaintiffs have proven the primacy effect
exists and affects Florida’s elections. This Court finds they have done so. The
second issue is whether the Constitution allows a state to put its thumb on the scale
and award an electoral advantage to the party in power. The answer is simple. It
does not.
https://www.ellsworthamerican.com/living/arts-a-living/dont-be-a-smartass-and-confusedonkeys-mules/ (explaining the history of the term “jackass,” meaning a male donkey).
2
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 3 of 74
The Florida Statute at Issue
Florida’s ballots are arranged in an office block pattern, meaning that all the
candidates for a given office are listed together in a section of the ballot labelled with
the title of the office they are running for. § 101.151(2)(a), Fla. Stat. (2019). Within
each office block, the candidates’ names are arranged according to the following
scheme:
The names of the candidates of the party that received the highest
number of votes for Governor in the last election in which a Governor
was elected shall be placed first for each office on the general election
ballot, together with an appropriate abbreviation of the party name; the
names of the candidates of the party that received the second highest
vote for Governor shall be placed second for each office, together with
an appropriate abbreviation of the party name.
Id. § 101.151(3)(a). 3 For all partisan races in the general election, therefore, the
candidates affiliated with the political party of the last-elected governor will be listed
first within each office block.
3
Florida law further provides that names of “[m]inor political party candidates” appear on
the ballot after the names of candidates of “recognized political parties” in the order they were
qualified to appear on the ballot. § 101.151(3)(b), Fla. Stat. (2019). Candidates with no party
affiliation appear last, in the order in which they were qualified. Id. Minor political parties are
those parties which have fewer than five percent of the total registered electorate as members.
§ 97.021(19), Fla. Stat. (2019). The statute does not explain what would result if a candidate from
a “minor political party” were elected governor or won the second-largest number of votes, or if
more than two parties were able to cross the five-percent threshold. The Florida Statutes do not
define “recognized political party.” Additionally, this ordering procedure applies only to general
elections. In primary elections, candidates are listed alphabetically within office blocks.
§ 101.151(4)(a), Fla. Stat. (2019).
3
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 4 of 74
Preliminary Miscellanea
In what has become a familiar exercise for this Court in cases concerning
voting rights and procedure, Defendants4 throw a hodgepodge of preliminary issues
at the wall, hoping one will stick and prevent this Court from considering this case
on the merits. See, e.g., Rivera Madera v. Detzner, 325 F. Supp. 3d 1269, 1275–78
(N.D. Fla. 2018); League of Women Voters of Fla., Inc. v. Detzner, 314 F. Supp. 3d
1205, 1212–14 (N.D. Fla. 2018); Fla. Democratic Party v. Detzner (Fla. Democratic
Party I), Case No. 4:16cv607-MW/CAS, 2016 WL 6090943, at *4–*5 (N.D. Fla.
Oct. 16, 2016); Fla. Democratic Party v. Scott (Fla. Democratic Party II), 215 F.
Supp. 3d 1250, 1254–55 (N.D. Fla. 2016). In the present case, Defendants claim
Plaintiffs lack standing; that Plaintiffs’ claims are barred by the applicable statute of
limitations, estoppel, and laches; and that this case is not justiciable.
In a prior order, this Court summarily rejected Defendants’ arguments on
these preliminary matters as “unpersuasive.” ECF No. 158 at 1. Defendants
renewed these arguments at trial citing supplemental authority, and this Court
remains unpersuaded. Although it entails a lengthy diversion to tilt at Defendants’
windmills, this Court will address each preliminary matter in turn.
4
For rhetorical convenience, this Court will refer to Defendant Lee and Intervenors
collectively as “Defendants.”
4
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 5 of 74
Justiciability
This Court need not struggle with the question of justiciability. The Supreme
Court has summarily affirmed a district court’s decision which held that, where
applicable law required candidates to be listed on the ballot in the order in which
they filed their qualification paperwork, a discretionary policy of resolving ties in
favor of incumbents was “a purposeful and unlawful invasion of [the] plaintiffs’
Fourteenth Amendment right to fair and evenhanded treatment.” Mann v. Powell,
314 F. Supp. 677, 679 (N.D. Ill. 1969), aff’d without opinion, 398 U.S. 988 (1970).
The Supreme Court’s summary affirmances are binding precedent unless and until
the Court specifically disclaims them. See Hicks v. Miranda, 422 U.S. 332, 344
(1975) (holding “that the lower courts are bound by summary decisions by this Court
until such time as this Court informs them that they are not” (internal marks and
quotation omitted)); see also Hardwick v. Bowers, 760 F.2d 1202, 1207 (11th Cir.
1985), rev’d on other grounds, Bowers v. Hardwick, 478 U.S. 186 (1986) (“A
summary affirmance of the Supreme Court has binding precedential effect.”).
The strength of this principle is such that summary affirmances by the
Supreme Court remain binding even in the face of decades of seemingly contrary
decisions. In Hand v. Scott, 285 F. Supp. 3d 1289 (N.D. Fla. 2018), this Court
addressed the precedential effect of the Supreme Court’s summary affirmance of
Beacham v. Braterman, 300 F. Supp. 182 (S.D. Fla. 1969), aff’d without opinion,
5
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 6 of 74
369 U.S. 12 (1969). This Court concluded it was not bound by Beacham because
“[u]nlike a fine wine, this summary affirmance has not aged well” in light of
subsequent Supreme Court decisions which seemed to recede from, or at least
contradict, that decision. 285 F. Supp. 3d at 1307. On appeal from this Court, the
United States Court of Appeals for the Eleventh Circuit admonished that “we are
bound to follow Supreme Court precedent in Beacham” and the Supreme Court’s
other “summary determinations.” Hand v. Scott, 888 F.3d 1206, 1208 (11th Cir.
2018). The summary affirmance of Mann would alone compel the conclusion that
Plaintiffs’ claims are justiciable.
Furthermore, in Cook v. Gralike, 531 U.S. 510 (2001), the Supreme Court
held a provision of the Missouri Constitution which required the words
“DISREGARDED VOTERS’ INSTRUCTIONS ON TERM LIMITS” to appear
alongside the names of certain candidates on Missouri’s ballots was
unconstitutional. Defendants agree Cook is binding authority—in fact, Intervenors
cited it during trial of this case in response to this Court’s question of whether a law
could be challenged which placed a “thumbs-up” symbol next to certain candidates
on the ballot, T. at 42–43, 70.5 Cook clearly holds that questions of what appears
5
Citations to the trial transcript will be formatted “T. at ___” followed by the transcript
page number.
6
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 7 of 74
on the ballot and how are justiciable. Defendants’ arguments make up in temerity
for what they lack in merit.
Notwithstanding those binding precedents, Defendants argue that the ordering
of candidates’ names on the ballot is a nonjusticiable political question, relying on
the Supreme Court’s recent decision in Rucho v. Common Cause, 139 S. Ct. 2484
(2019). In that case, the Court decided “that partisan gerrymandering claims present
political questions beyond the reach of the federal courts” because resolving such
claims would require federal courts “to reallocate political power between the two
major political parties, with no plausible grant of authority in the Constitution, and
no legal standards to limit and direct their decisions.” Id. at 2506–07. Such claims
are problematic, the Court explained, because they
inevitably ask the courts to make their own political judgment about
how much representation particular political parties deserve—based on
the votes of their supporters—and to rearrange the challenged districts
to achieve that end.
Id. at 2499. But the Court specifically distinguished other types of voting-rights
claims, such as “one-person, one-vote rule” claims which are “relatively easy to
administer as a matter of math,” id. at 2501; and even other types of gerrymandering
claims, such as racial gerrymandering, id. at 2502. Unlike claims of partisan
gerrymandering, the Court concluded, claims of vote dilution or racial
gerrymandering “could be decided under basic equal protection principles,” id. at
2496, without the need for a federal court to decide whether the legislature’s decision
7
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 8 of 74
is fair. Defendants contend that, if partisan gerrymandering claims are political
questions, ballot order claims must be too.
Hogwash. The legislative power is not a Midas touch that gilds a matter on
contact and insulates it from judicial review, and a decision does not become a
political question merely because it is made by a political branch of government.
See Fla. Democratic Party II, 215 F. Supp. 3d at 1258–59 (“It has been suggested
that the issue of extending the voter registration deadline is about politics.
Poppycock. This case is about the right of aspiring eligible voters to register and
have their votes counted.”). If this were so, federal courts would be unable to review
legislative enactments of any kind and would instead be bound to slavishly defer to
legislative enactments without respect to the dictates of the Constitution. This
plainly is not the case. 6 See also Williams v. Rhodes, 393 U.S. 23, 28 (1968)
(explaining the argument that ballot access claims were nonjusticiable political
questions had been “squarely rejected” and “requires little discussion”). And by its
plain terms, Rucho is an extremely narrow decision. The Court in Rucho explained
in no uncertain terms that its reasoning was limited to claims of partisan
gerrymandering, and did not even extend it to other types of gerrymandering claims,
6
The federal courts’ ability to review the constitutionality of legislative acts was not always
taken for granted but is now a well-established pillar of constitutional doctrine. See Marbury v.
Madison, 1 Cranch 137, 177–80 (1803) (elucidating the constitutional basis for judicial review of
legislative enactments and the supremacy of the Constitution); U.S. Const., art. VI, cl. 2
(establishing the Constitution as “the supreme Law of the Land”).
8
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 9 of 74
let alone other species of election law claims such as the one this case presents.
Rucho, 139 S. Ct. at 2501–02. 7
This case asks this Court to apply nothing more than “basic equal protection
principles,” id. at 2496, and is therefore justiciable under any fair reading of Rucho.
Plaintiffs do not ask this Court to determine whether section 101.151(3)(a) treats all
candidates fairly in light of their partisan affiliation; rather, they ask this Court to
determine whether it treats all candidates equally without regard to their partisan
affiliation. This question does not implicate issues of political “fairness,” and to
contend otherwise “is worse than solemn mockery.” See Marbury, 1 Cranch at 180.
As explained in more detail below, the issues presented in this case are not
novel. In a jurisprudential sense, they are not even particularly challenging. They
are “grounded in ‘a limited and precise rationale’ and [are] ‘clear, manageable, and
politically neutral,’ ” because they are “basic equal protection principles,” nothing
more. Rucho, 139 S. Ct. at 2496–98 (quoting Vieth v. Jubelirer, 541 U.S. 267, 306–
08 (2004) (Kennedy, J., concurring in the judgment)). The fact that this Court must
determine which level of scrutiny to apply in the context of an equal protection claim
also is not a new idea; in point of fact, Disney World, the states of Alaska and
7
Additionally, the Supreme Court “does not normally overturn, or so dramatically limit,
earlier authority sub silentio.” Shalala v. Ill. Council on Long Term Care, Inc., 529 U.S. 1, 18
(2000). Defendants’ arguments on justiciability ask this Court to conclude the Supreme Court in
Rucho did just that, and in dramatic fashion, receding from decades of decisions without even the
smallest explicit signal of that intention. This Court declines Defendants’ invitation to ignore this
foundational principle of stare decisis.
9
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 10 of 74
Hawaii, and the undersigned himself are newer additions to the world than this
concept.8 See, e.g., United States v. Carolene Prods. Co., 304 U.S. 144, 152 n.4
(1938) (noting the possibility of “a correspondingly more searching judicial inquiry”
and “more exacting judicial scrutiny under the general prohibitions of the Fourteenth
Amendment” when legislation impacts fundamental rights or protected classes). As
detailed below, the analytical framework this Court applies today has been applied
numerous times by this Court and other federal courts, without difficulty or
confusion. Cf. Rucho, 139 S. Ct. at 2491 (holding a claim was not justiciable
because, in part, courts had “struggled without success over the past several decades
to discern judicially manageable standards for deciding such claims”). Although
this Court must make hard judgments about the effects and justifications of the
challenged statute, there is no strange misprision in the law which might cloud this
Court’s perceptions and confuse its judgment. This Court simply does not face the
same sort of conundrum the Supreme Court did in Rucho.
In sum, therefore, Defendants claim this issue is not justiciable, but then rely
upon cases which are binding precedent on this Court and show not just that these
issues are indeed justiciable but also that the content and ordering of candidates on
8
The application of different levels of scrutiny to equal protection claims is very nearly
older than sliced bread. See Jim McCarty, Chillicothe’s Slice of History, City of Chillicothe, MO
(Sept. 2006) http://www.chillicothecity.org/bread/breadnews1.html (explaining sliced bread was
invented in 1928).
10
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 11 of 74
the ballot can violate voters’ constitutional rights. Defendants’ reliance on these
cases cannot be reconciled with their position that this Court cannot decide the
present case. In the alternative, Defendants ask this Court to transmogrify a recent
Supreme Court decision into a far more expansive ruling than it was, in contradiction
of clear, explicit limits announced in that decision itself and of its fundamental
rationale. Even if this Court had the power to overlook those restrictions, this Court
would nevertheless conclude this case was justiciable. No matter the trappings
Defendants try to drape it in, this donkey is a donkey, not a racehorse.
Standing
A party has standing to sue if they have suffered an injury in fact which is
fairly traceable to the defendant’s conduct and which is likely to be redressed by a
decision in their favor. Lujan v. Defenders of Wildlife, 504 U.S. 555, 560–61 (1992).
Defendants claim Plaintiffs have failed to prove the first and third elements of this
rule—that is, they claim Plaintiffs have not suffered an injury in fact and have not
proven their claims are redressable. Neither of these arguments passes muster.
Defendants’ arguments concerning injury in fact fall into two basic categories.
Regarding the organizational Plaintiffs in this action, Defendants contend those
Plaintiffs have not shown they suffered any concrete and particularized injury
because they did not show Florida’s ballot order statute impacted any specific
resource allocation decision; and Defendants furthermore contend that harm to
11
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 12 of 74
Democratic voters generally (as Defendants characterize Plaintiffs’ claims) is not
sufficient to impart standing to the organizational Plaintiffs.
Regarding the
individual Plaintiffs in this case, Defendants argue no individual Plaintiff proved
Florida’s ballot order statute impacted their individual voting rights—that is,
Florida’s ballot order statute did not prevent them from voting and having their vote
counted.
The organizational Plaintiffs have standing to sue based on injuries to
themselves or to their members, if either has been affected in a tangible way. See U.
Food & Commercial Workers Union Local 751 v. Brown Grp., Inc., 517 U.S. 544
(1996). An organization has standing to sue to
enforce the rights of its members “when its members would otherwise
have standing to sue in their own right, the interests at stake are
germane to the organization’s purpose, and neither the claim asserted
nor the relief requested requires the participation of individual members
in the lawsuit.”
Arcia v. Fla. Sec’y of State, 772 F.3d 1335, 1342 (11th Cir. 2014) (quoting Friends
of the Earth, Inc. v. Laidlaw Envt’l Servs. (TOC), Inc., 528 U.S. 167, 181 (2000)).
“[T]he rule in this Circuit is that organizational plaintiffs need only establish that ‘at
least one member faces a realistic danger’ of suffering an injury.” Id. (quoting Fla.
State Conference of the NAACP v. Browning, 522 F.3d 1153, 1163 (11th Cir. 2008)).
As this Court has previously explained, “political parties have standing to assert, at
least, the rights of [their] members who will vote in an upcoming election.” Fla.
12
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 13 of 74
Democratic Party II, 215 F. Supp. 3d. at 1254 (citing Fla. Democratic Party v. Hood,
342 F. Supp. 2d 1073, 1078–79 (N.D. Fla. 2004) (Hinkle, J.)). In that case, which
concerned extension of Florida’s voter registration period, this Court concluded the
plaintiff had standing to sue on its members’ behalf, reasoning the plaintiff “need
not identify specific aspiring eligible voters who intend to register as Democrats and
who will be barred from voting; it is sufficient that some inevitably will.” Id.
Plaintiffs in the present case go one step further.
Not only do the
organizational Plaintiffs provide evidence of how Florida’s ballot order statute
impacts their own interests and the interests of their numerous members, they are
also joined in this suit by individual Plaintiffs whose individual interests are affected
by Florida’s ballot order statute independent of their membership in an organization.
The organizational Plaintiffs represent their members, many of whom are voters, and
their common mission in one form or another is to support Democratic candidates
for elected office. Contrary to Defendants’ theory, an injury is not converted into a
“generalized grievance” solely because a large number of people are aggrieved by
it. Furthermore, “[e]ach provision of a[n election] code, ‘whether it governs the
registration and qualifications of voters, the selection and eligibility of candidates,
or the voting process itself, inevitably affects—at least to some degree—the
individual’s right to vote and his right to associate with others for political ends.’ ”
Burdick v. Takushi, 504 U.S. 428, 433 (1992) (quoting Anderson v. Celebrezze, 460
13
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 14 of 74
U.S. 780, 788 (1983)); see also id. at 434 (“There is no doubt that the Hawaii election
laws, like all election regulations, have an impact on the right to vote.” (emphasis
added)). In contrast to Defendants’ hyper-formalistic approach to voting rights, “the
rights of voters and the rights of candidates do not lend themselves to neat
separation; laws that affect candidates always have at least some theoretical,
correlative effect on voters.” Bullock v. Carter, 405 U.S. 134, 143 (1972) (emphasis
added). It is thus no answer to Plaintiffs’ claims for Defendants to argue that
individual Plaintiffs have been able to vote and have their vote counted, or that
organizational Plaintiffs do not vote but instead attempt to support and elect
Democratic candidates. 9 This Court declines to adopt Defendants’ cramped (and
jurisprudentially anomalous) interpretation of the scope of the right to vote. 10
It is worth noting again that, in seeking to intervene in this case, Intervenors
claimed Republican candidates and the organizations which support them—as
organizational Plaintiffs support Democratic candidates—“stand to be most directly
9
Furthermore, drain on resources has long been recognized as sufficient grounds for
organizational standing. Havens Realty Corp. v. Coleman, 455 U.S. 363, 379 (1982). The
testimony presented during trial was more than sufficient to establish that organizational Plaintiffs
have standing under this theory as well.
10
Intervenors note that Plaintiffs Bottcher and Fleming did not testify at trial and were not
produced for de bene esse depositions. Therefore, Intervenors claim, Plaintiffs Bottcher and
Fleming lack standing. This Court disagrees. The cases cited in this paragraph—precedent this
Court is bound to follow—do not limit the “impact on the right to vote” common to all election
laws, Burdick, 504 U.S. at 433, to a discrete group or subclass of voters. Plaintiffs Bottcher and
Fleming introduced sufficient evidence to establish their individual standing, although they
themselves did not testify.
14
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 15 of 74
harmed by a change” in Florida’s ballot order scheme. ECF No. 23 at 16. In other
words, Intervenors’ whole interest in this lawsuit rests upon the idea that a change
in Florida’s ballot order scheme would worsen their position by depriving them of
an existing advantage. But they now claim that Florida’s ballot order scheme does
not injure Plaintiffs, or in other words that Democratic candidates suffer no
cognizable disadvantage, and that Plaintiffs therefore lack standing. Because the
electorate is finite, elections are a zero-sum game, and this Court is at a loss to see
how awarding an advantage to one group of candidates would not also disadvantage
a competing group. See Green Party of Tenn. v. Hargett, 791 F.3d 684, 695 (6th
Cir. 2015) (holding that requiring minor political parties to pass same electoral
threshold as major parties in less time violates the Equal Protection Clause); Nat.
Law Party of U.S. v. Fed. Election Comm’n, 111 F. Supp. 2d 33, 44 (D.D.C. 2000)
(explaining that, in the context of Equal Protection Clause claims, injury arises from
denial of the opportunity to compete on an equal footing). Intervenors cannot
contend both that their interests are implicated but Plaintiffs have not been injured
without doing violence to basic logic. Plaintiffs have satisfied the first element of
Article III standing.
Defendants also claim Plaintiffs have not satisfied the redressability prong of
standing; first, because Plaintiffs’ requested solution will not remove the influence
vel non of the primacy effect on elections, and second, because this Court lacks the
15
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 16 of 74
power to order a specific remedy. These two distinct arguments share a common
flaw in that they misconstrue Plaintiffs’ claims. Plaintiffs’ alleged injury in this case
is not based on the mere existence of the primacy effect vote and its impact on
elections. Rather, their claims concern the fact that Florida’s ballot order statute
allocates the primacy effect vote to groups of candidates on the sole basis of partisan
affiliation. If the remedy (or remedies) Plaintiffs seek would cure this alleged injury
on either basis—that is, by eliminating the primacy effect vote altogether, or by
removing the alleged partisan basis for its allocation—then Plaintiffs have shown
redressability. In addition, “[t]o have Article III standing, a plaintiff “need not
demonstrate anything ‘more than . . . a substantial likelihood’ of redressability.”
Wilding v. DNC Servs. Corp., No. 17-14194, 2019 WL 5539021, at *5 (11th Cir.
Oct. 28, 2019) (quoting Duke Power Co. v. Carolina Envt’l Study Grp., Inc., 438
U.S. 59, 79 (1978)) (alteration in original); see also id. (noting “even partial relief
suffices for redressability” and citing Made in the USA Found. v. United States, 242
F.3d 1300, 1310–11 (11th Cir. 2001)). Defendants’ contention that Plaintiffs have
failed to identify a complete solution to the claimed injury is therefore inapposite.
Furthermore, this Court need not have the ability to order a specific remedy
for Plaintiffs to satisfy the redressability element of standing, so long as this Court
has the power to order a remedy which will to some extent cure Plaintiffs’ injury.
If, as Plaintiffs allege, there are multiple means of arranging the candidates on
16
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 17 of 74
Florida’s ballots which would satisfy the Constitution and prevent Plaintiffs’ alleged
injuries from continuing forward, all that is required for Plaintiffs to have standing
is for this Court to be empowered—as it is empowered—to say what the Constitution
does and does not permit, even if this Court cannot order the Florida Legislature to
adopt a specific ballot order scheme from among the available options. Indeed, as
explained below, this is precisely what this Court will do. As a practical matter,
there must be some sort of ballot order scheme, and as this Court will explain, the
choice of which ballot order scheme Florida shall use is in this case a question for
the Florida Legislature to resolve. But this Court has the power to determine whether
Florida’s existing ballot order scheme comports with the Constitution and to explain
what the Constitution does and does not allow, i.e., “to say what the law is,”
Marbury, 1 Cranch at 177, and thus to decide whether there are certain ballot order
schemes Florida cannot use. “This is of the very essence of judicial duty.” Id. at
178.
Finally, Intervenors claim this Court cannot redress Plaintiffs’ injuries
because any new ballot order scheme would have to be implemented by third parties
not before this Court; i.e., by county supervisors of elections and not Defendant Lee.
If this argument sounds familiar to readers well versed in this Court’s voting rights
jurisprudence, there is a good reason for that. We have been here before.
As this Court notes with tiresome regularity, Defendant [Lee] is
Florida’s “chief election officer.” This statutory job description is not
17
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 18 of 74
window dressing. The Secretary of State must “obtain and maintain
uniformity in the interpretation and implementation of the election
laws” and promulgate rules for the “proper and equitable interpretation
and implementation of election laws. Additionally, the Department of
State “shall have general supervision and administration of the election
laws, corporation laws and such other laws as are placed under it by the
Legislature.” . . . Defendant [Lee] has the responsibility to enforce the
Department of State’s rules on each county supervisor of elections.
[She] also has the responsibility to “provide written directions and
opinions to the supervisors of elections on the performance of their
official duties with respect to . . . rules adopted by the Department of
State.”
Rivera Madera, 325 F. Supp. 3d at 1276 (internal marks and citations omitted) (final
alteration in original). The Eleventh Circuit has agreed, denying a stay of this
Court’s ruling in a different case on the basis that, “[b]ecause the Secretary is the
state’s chief election officer with the authority to relieve the burden on Plaintiff’s
right to vote, she was appropriately sued for prospective injunctive relief.”
Democratic Exec. Comm. of Fla. v. Lee, 915 F.3d 1312, 1318 (11th Cir. 2019). This
Court, the parties, and—pursuant to the prior panel rule—even the Eleventh Circuit
itself are bound by this holding unless and until the Eleventh Circuit recedes from it
en banc. This issue is settled. It is perplexing that Intervenors would continue to
raise it. Defendant Lee is the proper defendant against whom to award the relief
Plaintiffs seek.
Plaintiffs have shown they have suffered an injury in fact and this Court has
the power to vindicate Plaintiffs’ claims. Plaintiffs therefore have standing.
18
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 19 of 74
Statute of Limitations
Next, Defendant Lee claims Florida’s statute of limitations bars Plaintiffs’
claims. 11 Florida law provides a four-year limitation period for “action[s] founded
on statutory liability,” § 95.11(3)(f), Fla. Stat. (2019), and for “[a]ny action not
specifically provided for in these statutes,” id. § 95.11(3)(p). Florida’s statute of
limitations applies to the claims at issue in this action. See Boyd v. Warden, Holman
Corr. Facility, 856 F.3d 853, 872 (11th Cir. 2017) (“All constitutional claims
brought under § 1983 are tort actions and, thus, are subject to the statute of
limitations governing personal injury actions in the state where the § 1983 action has
been brought.”). In Florida, “[a] cause of action accrues when the last element
constituting the cause of action occurs.” § 95.031(1), Fla. Stat. (2019); see also
Hearndon v. Graham, 767 So. 2d 1179, 1184–85 (Fla. 2000). Generally, the last
element which will occur is the injury giving rise to the claim. State Farm Mut.
Auto. Ins. Co. v. Lee, 678 So. 2d 818, 821 (Fla. 1996).
The Florida Legislature adopted Florida’s ballot order scheme in 1951.
Defendant Lee claims all subsequent harms flow from this enactment and therefore
the statute of limitations began to run at that point, meaning Plaintiffs’ timely filing
11
Plaintiffs argue Defendant Lee waived this argument. Plaintiffs are mistaken. It is true
Defendant Lee raised the statute of limitations defense for the first time in her motion for summary
judgment, ECF No. 115 at 32, but Eleventh Circuit precedent allows this so long as the opposing
party has sufficient notice of the defense and a chance to rebut it. Navarro v. Santos Furniture
Custom Design, Inc., 372 F. App’x 24, 27 (11th Cir. 2010). Defendant Lee therefore did not waive
this defense.
19
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 20 of 74
window expired in 1955.12 Of course, the issue is more complicated than that, and
ultimately ties itself in knots. Each individual Plaintiff became aware of the alleged
existence and effects of the primacy effect vote at various times, though under
Florida law it is not clear whether this would be relevant. Compare Davis v.
Monahan, 832 So. 2d 708, 710–12 (Fla. 2002) (declining to apply the delayed
discovery doctrine absent a statutory basis or allegations that the defendant’s
conduct delayed the plaintiff’s discovery of the factual basis for the lawsuit) with
Chappell v. Rich, 340 F.3d 1279, 1283 (11th Cir. 2003) (delayed discovery doctrine
applies to section 1983 claims). And if any Plaintiff was not born (or founded) until
after that point, what of their claim’s accrual date? Assuming no other issues of
access to courts apply, 13 under Defendant Lee’s theory, Plaintiffs’ individual claims
accrued either in 1951 or at whatever point thereafter their constitutional rights were
first allegedly violated, and became time-barred four years thereafter.
12
In the most technical sense, the statute of limitations applicable to Plaintiffs’ claims
under Defendant Lee’s theory of the case is not the 2018 version (current when Plaintiffs filed this
action) nor the 2019 version (current now), but rather the 1951 version. See Merkle v. Robinson,
737 So. 2d 540, 542 n.6 (Fla. 1999) (statutes of limitation “bar actions by setting a time limit within
which an action must be filed as measured from the accrual of the cause of action”); Dade Cty. v.
Rohr Inds., Inc., 826 F.2d 983, 989 (11th Cir. 1987) (ruling “that the timing of the cause of action
determines the applicable statute of limitations”). The 1951 version also provided a four-year
limitations period. See § 95.11(4), Fla. Stat. (1951). For uniformity and convenience, this Court
will cite the 2019 Florida Statutes throughout this Order, except where some relevant difference
exists between the applicable version and current version beyond the numbering of the sections
and subsections.
13
See, e.g., Cates v. Graham, 451 So. 2d 475, 476–77 (1984) (discussing the interplay
between statutes of limitation, statutes of repose, and the Florida Constitution’s right of access to
courts).
20
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 21 of 74
If section 101.151(3)(a), Florida Statutes, had been applied only to an election
seventy years ago, Defendant Lee’s argument might have merit. As things stand,
however, irrespective of how many times Plaintiffs allege their rights have been
violated in the past, their rights are violated anew each time an election is held and
the candidates’ names are arranged in their respective office blocks in the order
prescribed by section 101.151(3)(a). Moreover, Plaintiffs seek declaratory and
injunctive relief, not damages or other retrospective relief. By way of analogy,
assume there was a Florida Statute which required all registered Republicans to pay
$10 at their designated polling site before voting in each election. Assume this
flagrantly illegal statute was adopted in 1910, but never challenged—instead, for
more than one hundred years, all registered Republicans quietly paid the infamous
poll tax. Further assume that, in 2018, a few Republican voters—backed by
Republican political organizations—decided enough was enough, and it was time to
strike that statute from the books so as to avoid paying the $10 during the 2020
election. Each historical $10 payment would be a separate, recurring injury.
In the present case, Plaintiffs contend their rights were violated each time an
election was held and the challenged statute was enforced. The most recent election
before Plaintiffs filed their claims was in 2016, and Plaintiffs filed their claims in
2018. If Florida’s ballot order statute stands, it will be enforced during the 2020
election cycle and, Plaintiffs allege, their injuries will recur. Plaintiffs’ claims relate
21
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 22 of 74
to alleged violations of their rights which have occurred in the recent past and will
continue to recur in the future. See Hillcrest Prop., LLC v. Pasco Cty., 754 F.3d
1279, 1282–83 (11th Cir. 2014) (explaining that, outside the context of substantive
due process takings claims, “the harm inflicted by the statute is continuing, or does
not occur until the statute is enforced—in other words, until it is applied,” quoting
Levald, Inc. v. City of Palm Desert, 998 F.2d 680, 688 (9th Cir. 1993)); see also
Moore v. Ogilvie, 394 U.S. 814, 816 (1969) (holding election statute unconstitutional
although the challenged election was over because the statute “remain[ed] and
control[led] future elections,” and the issue was therefore not moot). In short, the
statute of limitations does not bar Plaintiffs’ claims.
Laches and Constitutional Estoppel
Finally, Defendants each argue Plaintiffs’ claims are barred by separate
equitable doctrines. Specifically, Intervenors assert the doctrine of laches bars
Plaintiffs’ claims, and Defendant Lee asserts the doctrine of estoppel—to be precise,
constitutional estoppel—bars them. Neither argument has merit.
To succeed on a laches claim, a defendant must demonstrate the plaintiff
inexcusably delayed in bringing their claim and that such delay unduly prejudiced
the defendant. United States v. Barfield, 396 F.3d 1144, 1150 (11th Cir. 2005). It
is not clear whether laches applies at all to claims for prospective relief from
continuing constitutional violations. See Garza v. Cty. of Los Angeles, 918 F.2d 763,
22
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 23 of 74
772 (9th Cir. 1990); Peter Letterese & Assocs., Inc. v. World Inst. of Scientology
Enters. Int’l, 533 F.3d 1287, 1321 (11th Cir. 2008) (stating, in a copyright case, that
“laches serves as a bar only to the recovery of retrospective damage, not to
prospective relief”); see also Democratic Exec. Comm. of Fla., 915 F. 3d at 1326
(“We need not consider whether laches applies to bar prospective relief from
constitutional harms, because the NRSC cannot satisfy the laches elements.”).14
This Court would conclude laches does not apply to the kind of prospective relief at
issue in this case; but even if it does, Intervenors’ laches claim fails on the merits.
First, Plaintiffs have not inexcusably delayed bringing their claims because
the event chiefly complained of—the 2020 election—has not happened yet.
Whether Plaintiffs could have made similar claims about an earlier election is not
relevant. Plaintiffs do not seek redress for past harms, they seek prospective relief
from future ones. Furthermore, it is unclear whether Plaintiffs would have been able
to prosecute their claims without substantial data on the primacy effect collected
over time—a fact which would arguably excuse at least some delay. Second, even
assuming Plaintiffs had delayed without a valid excuse, Intervenors have not shown
they have suffered any undue prejudice. The undue prejudice cognizable in laches
14
It is also not clear that laches can bar a claim where the applicable statute of limitations
has not run. See Merck & Co. v. Reynolds, 559 U.S. 633, 652 (2010) (“Laches within the term of
the statute of limitations is no defense at law.”) (quoting United States v. Mack, 295 U.S. 480, 489
(1935)).
23
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 24 of 74
can be either substantive or procedural—that is, it may concern both the inability to
defend against claims because memory fades or evidence is otherwise unavailable,
and also the reliance costs of acting in conformity with what one believed the law to
be during the period of delay; but it does not embrace any delay or expenditure
whatsoever. See, e.g., Gardner v. Panama R.R. Co., 342 U.S. 29, 30 (1951) (per
curiam) (explaining laches will not bar relief “where no prejudice to the defendant
has ensued from the mere passage of time”); Black Warrior Riverkeeper, Inc. v. U.S.
Army Corps of Eng’rs, 781 F.3d 1271, 1286 (11th Cir. 2015) (“Any harm
demonstrated by the Intervenors must stem specially from Riverkeeper’s delay in
bringing suit, rather than from the consequences of an adverse decision on the merits
. . .. The paradigmatic example of prejudice is when a defendant has expended
substantial sums of money or completed a significant amount of construction by the
time the plaintiff decides to file suit.”). The only prejudice Intervenors identify is
the cost to the State of Florida in time and resources to adjust its conduct of elections
into conformity with the Constitution, if this Court finds for Plaintiffs. This is not
equivalent in kind or quantity to, say, demolishing a building. It is not a cognizable
prejudice within the doctrine of laches at all, much less an undue one—it is precisely
and exclusively a “consequence[] of an adverse decision on the merits.”
Id.
Intervenors might be able to claim prejudice if Plaintiffs had sought, for instance, to
24
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 25 of 74
invalidate past results and hold new elections; but Plaintiffs do not. Even assuming
the doctrine of laches could bar this action, Intervenors’ argument fails on the merits.
Turning to constitutional estoppel, Defendant Lee contends Plaintiffs cannot
argue against the constitutionality of Florida’s ballot order scheme because, in the
past, Democrats have been governors of Florida and Democratic candidates have,
fittingly, benefitted from the donkey vote. In general, constitutional estoppel applies
to prevent a party from enjoying the benefits of a government action while
simultaneously challenging the constitutionality of that government action.
Robertson v. Fed. Election Comm’n, 45 F.3d 486, 489–90 (D.C. Cir. 1995). But
“[t]he doctrine of constitutional estoppel . . . has its limits. The government may not
interpose the doctrine as a defense if a party wishes to challenge an unconstitutional
condition which is imposed on the receipt of federal funds.” Id. at 490 (citing
Kadrmas v. Dickinson Pub. Sch., 487 U.S. 450, 456–57 (1988)). And the Supreme
Court has expressed “doubt that plaintiffs are generally forbidden to challenge a
statute simply because they are deriving some benefit from it.” Kadrmas, 487 U.S.
at 456–57. Plaintiffs in this case are not seeking to, for example, retain federal funds
awarded by an agency while also challenging the existence or composition of that
agency as unconstitutional. Cf. Robertson, 45 F.3d at 488–89. And whatever the
situation was in the past, Democratic candidates certainly do not enjoy any
advantage or benefit as a result of Florida’s ballot scheme at present. The doctrine
25
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 26 of 74
does not apply here at all, and Plaintiffs are not constitutionally estopped from
pursuing their claims in this case.
Having disposed of those preliminary matters, this Court will now analyze the
merits of the case.
Merits of the Case
“Voting is the beating heart of democracy. It is a ‘fundamental political right,
because [it is] preservative of all rights.’ ‘It is beyond cavil that voting is of the most
fundamental significance under our constitutional structure.’ ” League of Women
Voters, 314 F. Supp. 3d at 1215 (alteration in original) (citations omitted). As the
Eleventh Circuit recently opined: “We can’t say it any better than that.” Democratic
Exec. Comm., 915 F.3d at 1315. And it is central to the civic life of our Republic
that elections shall not just be held—because “voting alone is not enough to keep
democracy’s heart beating,” id.—but that they shall be fair and free, and that the
state shall not take sides. We have neither monarch, nor nobles, nor electors palatine,
and in our democracy there are no head starts.
“It does not follow, however, that the right to vote in any manner and the right
to associate for political purposes through the ballot are absolute.” Burdick, 504
U.S. at 433. “[A]s a practical matter, there must be substantial regulation of elections
if they are to be fair and honest and if some sort of order, rather than chaos, is to
accompany the democratic processes.” Storer v. Brown, 415 U.S. 724, 730 (1974).
26
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 27 of 74
That said, “these granted powers are always subject to the limitation that they may
not be exercised in a way that violates other specific provisions of the Constitution.”
Williams, 393 U.S. at 29. To evaluate First and Fourteenth Amendment voting
claims, a federal court
must first consider the character and magnitude of the asserted injury
to the rights protected by the First and Fourteenth Amendments that the
plaintiff seeks to vindicate. It must then identify and evaluate the
precise interests put forward by the State as justifications for the burden
imposed by its rule. In passing judgment, the Court must determine not
only the legitimacy and strength of each of those interests; it must also
consider the extent to which those interests make it necessary to burden
the plaintiff’s rights. Only after weighing all these factors is the
reviewing court in a position to decide whether the challenged
provision is unconstitutional.
Anderson, 460 U.S. at 789. “Under this standard, the rigorousness of [a reviewing
court’s] inquiry into the propriety of a state election law depends upon the extent to
which a challenged regulation burdens First and Fourteenth Amendment rights.”
Burdick, 504 U.S. at 434. If a law imposes a mild or minor burden on these rights,
it can usually be justified by showing the law in question serves a legitimate state
interest. See Anderson, 460 U.S. at 788 (“Nevertheless, the state’s important
regulatory interests are generally sufficient to justify reasonable, nondiscriminatory
restrictions.”). Although that level of scrutiny presents a low bar, it is not a
meaningless one, and “[h]owever slight that burden may appear, . . . it must be
justified by relevant and legitimate state interests ‘sufficiently weighty to justify the
limitation.’ ” Crawford v. Marion Cty. Election Bd., 553 U.S. 181, 191 (2008)
27
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 28 of 74
(quoting Norman v. Reed, 502 U.S. 279, 288–89 (1992)). “Laws imposing ‘severe’
burdens, on the other hand, ‘must be narrowly drawn to advance a state interest of
compelling importance.’ ” Libertarian Party of Va. v. Alcorn, 826 F.3d 708, 717
(4th Cir. 2016) (quoting Burdick, 504 U.S. at 434). In all cases, courts conduct a
“realistic appraisal” of the challenged law, Anderson, 460 U.S. at 806, and examine
whether the state’s asserted interests “justify the specific restriction . . . at issue,” id.
at 796.
As this Court has explained above, there is nothing novel about this
framework, at any level.
This Court and other courts have applied the
Anderson/Burdick interest-balancing framework to a wide variety of voting-related
issues, including claims relating to ballots. See Buckley v. Am. Constitutional Law
Found., Inc., 525 U.S. 182, 197–200 (1999) (evaluating Colorado statute requiring
those circulating initiative petitions to wear identification badges); Democratic Exec.
Comm., 915 F.3d at 1319 (evaluating Florida’s voter signature-matching scheme).
Courts have also applied it to cases involving ballot order. See Alcorn, 826 F.3d at
714–21; Graves v. McElderry, 946 F. Supp. 1569, 1579–82 (W.D. Okla. 1996). The
only thing distinguishing the present case from those cases is, the present case
represents the first time there has been a fully developed evidentiary record upon
which to evaluate a ballot order claim. And as this Court earlier explained, the
Supreme Court’s summary affirmance in Mann indicates ballot order claims
28
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 29 of 74
implicate the Equal Protection Clause, see 314 F. Supp. at 679, and that summary
affirmance is binding on this Court.
There is nothing new about the
Anderson/Burdick standard, nor its application to this type of claim.
To apply the Anderson/Burdick standard, this Court will first examine whether
and to what extent Plaintiffs’ rights have been burdened by Florida’s ballot order
scheme. This entails investigating whether the primacy effect exists, how large it is,
and the character of its effects on Plaintiffs’ rights. Next, this Court will consider
the justifications Defendants advance for Florida’s current ballot order scheme and
determine the strength and legitimacy of those justifications. Finally, this Court will
weigh the burden on Plaintiffs’ rights against Defendants’ justifications to determine
whether Florida’s current ballot order scheme violates Plaintiffs’ First and
Fourteenth Amendment rights.
The existence and magnitude of the primacy effect vote have been the central
questions of fact in this case. This Court denied the cross-motions for summary
judgment in this case because there were “material issues of disputed fact regarding
‘position bias’ and its purported effects on Florida elections as well as the viability
of alternatives to the current [ballot order] scheme.” ECF No. 158 at 1. Following
the contours of the Anderson/Burdick standard, the bench trial held in this case
focused on whether Florida’s ballot order scheme awarded candidates affiliated with
the party of the last-elected governor a systematic advantage on the basis of their
29
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 30 of 74
political affiliation. Defendants elected to center their defense on disputing whether
Plaintiffs have proven the quantum of the primacy effect, not on whether it exists at
all. Defendants did not call an expert to dispute the existence of the primacy effect.15
Instead, as will shortly be seen, their expert only disputed Plaintiffs’ experts’
calculations concerning the magnitude of the primacy effect.
This proceeding was a bench trial, and this Court therefore sits as the
factfinder. This Court must consider all the evidence presented, but it need not
accept all the evidence as true or accurate. This Court must decide whether it
believes what each witness had to say, and how important that testimony was. This
Court may believe or disbelieve any witness, in whole or in part, just as a jury would
be free to do. This includes expert witnesses: as with any other witness, this Court
must weigh the testimony of an expert witness and decide whether to rely upon that
expert’s opinion. Furthermore, this Court may use reasoning and common sense to
make deductions and reach conclusions, just as a jury would.
As explained below, this Court concludes section 101.151(3)(a) imposes a
discriminatory burden on Plaintiffs’ voting rights which is not of the same magnitude
as entirely denying Plaintiffs the franchise, but is not negligible either. This Court
15
The parties offered numerous studies on the primacy effect into evidence, and only one
of those—a study concerning elections in Afghanistan—found no statistically significant evidence
of a primacy effect. Pls. Ex. 1 at 29–32 (citing B.T. Hansen and A.L. Olsen, Order in chaos:
Ballot order effects in a post-conflict election?, Research and Politics 1–4 (Oct.-Dec. 2014)). This
lone exception, however, does not disprove the rule.
30
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 31 of 74
further finds many of the justifications Defendants offered are not justifications for
Florida’s specific ballot order scheme, but rather justifications for not changing an
existing ballot order scheme irrespective of what that scheme happens to be. To the
extent Defendants’ justifications pertain to Florida’s specific ballot order scheme,
this Court finds they are weak and entitled to little weight.
Weighing the
discriminatory burden on Plaintiffs’ rights against Defendants’ meager
justifications, this Court concludes Florida’s current ballot order scheme violates
Plaintiffs’ First and Fourteenth Amendment rights.
Character and Magnitude of Injury
The first step in this Court’s analysis is to “consider the character and
magnitude of the asserted injury to the rights protected by the First and Fourteenth
Amendments that the plaintiff seeks to vindicate.” Anderson, 460 U.S. at 789. The
character and magnitude of Plaintiffs’ injuries are functions of the primacy effect
itself—of whether it exists, how large it is, and what effect it has under Florida’s
current ballot order scheme. Consideration of these items determines the weight this
Court will accord to the burden on Plaintiffs’ rights, and it also calibrates the scales
upon which that weighing is done by determining the appropriate level of scrutiny
to which this Court will subject the challenged law. See Anderson, 460 U.S. at 788
(stating “the state’s important regulatory interests are generally sufficient to justify
reasonable, nondiscriminatory restrictions”); Burdick, 504 U.S. at 434 (explaining
31
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 32 of 74
laws which severely burden voting rights “must be narrowly drawn to advance a
state interest of compelling importance”).
As to magnitude, Dr. Jon Krosnick, a political scientist and a preeminent
expert on candidate name order effects,16 testified on Plaintiffs’ behalf. See T. at
279–90, 292–93. Although Dr. Krosnick’s extensive qualifications and status as the
preeminent expert in his field do not automatically establish his credibility, this
Court finds them probative of that issue. This Court does not blindly credit Dr.
Krosnick’s methods, analysis, and conclusions; but after close and attentive scrutiny,
this Court finds Dr. Krosnick’s methods and conclusions reasonable, reliable, and
credible.
Dr. Krosnick provided a comprehensive review of seventy years’ worth of
academic scholarship on candidate name order effects, which overwhelmingly found
primacy effects were evident in the elections studied across that time period. T. at
295–96, 302–25; Pls. Ex. 1 at 15–39. To be precise, of the 1,086 unique tests
reported in the literature, eighty-four percent manifested differences in the direction
of primacy, a result which Dr. Krosnick calculated had a less than one onethousandth of a percent chance (in fact, a 0.0001% chance) of occurring for some
16
Although this Court and numerous other sources refer to this phenomenon by many
names, such as the “primacy effect” or “ballot order effect,” Dr. Krosnick testified that “candidate
name order effects” is “the clearest term” to refer to “the possibility that the order in which
candidates’ names appear on a ballot might influence the behavior of voters and, therefore,
influence the outcome of an election.” T. at 290–91.
32
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 33 of 74
reason other than candidate name order. T. at 304. Dr. Krosnick also testified
concerning the effect of candidate name order on Florida elections from 1978
through 2016. T. at 299. Based on his study of these results, Dr. Krosnick opined
that first-listed candidates in Florida have historically gained an average advantage
of five percentage points 17 due to their position within their office block on the
ballot, T. at 299–301; Pls. Ex. 1 at 3, 63–64, 83; a result which has a less than onepercent probability of occurring by chance. T. at 343; Pls. Ex. 1 at 110. Moreover,
because votes for one candidate are necessarily votes which do not go to another,
Dr. Krosnick testified the five-percent advantage enjoyed by first-listed candidates
equates to a ten-percent difference between those candidates and their competitors
in two-candidate races. T. at 299.
To oppose Dr. Krosnick, Intervenors called Dr. Michael Barber, a political
science scholar and professor at Brigham Young University who specializes in
campaign finance and quantitative analysis. T. at 597–609. As noted above, Dr.
Barber did not dispute the central findings of the significant body of academic
literature concerning candidate name order effects, nor did he dispute that such
17
Dr. Krosnick testified Republican candidates in Florida have historically enjoyed an
average primacy effect advantage of approximately 5.35% while Democratic candidates in Florida
have historically received an average advantage of 4.57%. T. at 301. The average of these two
primacy effect advantages—and thus the approximate average for all candidates in partisan races
in Florida for the years studied (because addition and multiplication are, after all, commutative)—
is 4.96%, which Dr. Krosnick rounded up to an estimated average primacy effect in Florida of 5%.
Id. Although Dr. Krosnick’s off-the-cuff calculation of this average was 4.97%, id., this Court
does not find this minor discrepancy affects his credibility in a negative way.
33
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 34 of 74
effects are more pronounced in down-ballot races. T. at 711, 729. Instead, Dr.
Barber focused his testimony on speculative critiques of Dr. Krosnick’s methods
with particular relation to whether Dr. Krosnick’s estimate of a 5% average primacy
effect vote was reliable. Dr. Barber suggested Dr. Krosnick could have adjusted his
study to control for certain demographic characteristics that are linked to different
voting preferences, T. at 619–44, but clarified upon questioning by this Court that
such disaggregation was only appropriate “if you have a theoretical reason to believe
those differences are meaningful” to the hypothesis one is attempting to investigate.
T. at 624–25. But what Dr. Barber testified was that “it’s not unreasonable to
believe that it’s very possible that those differences could manifest themselves in
other ways related to elections, such as ballot order effects.” T. at 621 (emphasis
added). Indeed, this Court had sufficient opportunity to closely observe not just the
substance of Dr. Barber’s testimony, but his demeanor as a witness, and his labored
responses to questioning by counsel and this Court serve only to highlight his
unconvincing equivocations.
On cross-examination by Ms. Khanna for the Plaintiffs, Dr. Barber’s
testimony crumbled. In an artful tour de force that should be studied by law students
as a textbook demolition of an expert witness, Ms. Khanna methodically walked Dr.
Barber through his direct examination testimony and forced him to admit almost all
of it was either speculative, unsound, or both. Dr. Barber admitted that, although
34
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 35 of 74
Dr. Krosnick could have controlled for those demographic factors in his regression
analysis, there was no reason Dr. Krosnick should have, and none to believe doing
so would make a material difference—in other words, there was no reason to believe
those demographic variables had any meaningful impact on the validity of Dr.
Krosnick’s conclusions. T. at 713–19. In sum, Dr. Barber admitted that, although
the demographic factors he identified correlate with individuals’ preferences
expressed through voting, there is no reason to believe they are connected in any
way to voters’ susceptibility to candidate name order effects. In plain English,
although a person’s demographic characteristics can help predict how they may
consciously choose to vote, they do not predict their susceptibility to the
psychological phenomenon at issue in this case.
Dr. Barber’s testimony demonstrated that, for his analysis of this case,
statistical rigor was a matter of convenience, not principle. For example, Dr. Barber
criticized Dr. Krosnick’s analysis concerning its small sample size, but also tried to
undermine Dr. Krosnick’s conclusions by pointing to outlier results in individual
elections and attempting to generalize from them, a practice this Court finds
unreliable and nonsensical. In a further baffling inconsistency, although Dr. Barber
criticized Dr. Krosnick’s methodology and attempted to discredit his results, Dr.
Barber testified he did not dispute the numerous other studies in the record which
found and analyzed candidate name order effects. See, e.g., T. at 729. It is probative
35
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 36 of 74
of both Dr. Barber’s credibility and reliability, and also of Dr. Krosnick’s credibility
and reliability, that Dr. Krosnick’s analysis agrees with the overwhelming majority
of the academic literature in the record, and Dr. Barber does not dispute the
reliability of those other studies.
Ms. Khanna’s skillful evisceration of Dr. Barber’s testimony was sufficient to
undermine his conclusions, but—as Ms. Khanna also elucidated through her crossexamination—Dr. Krosnick also performed a second regression analysis
incorporating those of Dr. Barber’s suggestions that could have affected the analysis.
This second analysis returned a slightly lower estimate of an overall average primacy
effect—about three-and-a-half percent, all told. T. at 730–31, 740–42; Pls. Ex. 1 at
73–74, 78–80; Pls. Ex. 11 at 4–5, 10. This small change did not in any way
undermine Dr. Krosnick’s conclusion that candidate name order has a statistically
significant impact on Florida’s elections. Dr. Krosnick also followed Dr. Barber’s
suggestion and weighted counties by size, but this did not significantly impact his
results either. T. at 730–31; Pls. Ex. 1 at 73–74; Pls. Ex. 11 at 10.
The same was true of Dr. Barber’s suggestion that clustering of standard errors
could improve Dr. Krosnick’s analysis: Dr. Krosnick performed a second analysis
in which he clustered standard errors as Dr. Barber suggested, and that analysis
found candidate name order effects of a similar magnitude, though at a marginally
reduced level of certainty. T. at 666–67, 736–43. For example, clustering by
36
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 37 of 74
governor cycle reduced the p-value to .14 for Republican candidates and .16 for
Democratic candidates, reflecting an 84% and 86% chance, respectively, that the
increased vote share was a result of candidate name order effects. Id.
This does not undermine Dr. Krosnick’s results, for two reasons. First, an
84% or 86% level of confidence in a result is sufficient, on the facts of this case, for
this Court to find it credible, especially in concert with the totality of the testimony
presented in this case. 18 The burden of proof does not require Plaintiffs to establish
the existence and magnitude of candidate name order effects with absolute statistical
certainty—although in this case this Court finds they very nearly have—but rather
to establish them by a preponderance of the evidence. Second, as Dr. Krosnick
described, although clustering can be helpful, too much clustering can have the effect
of obscuring the detail of a result and making it appear less accurate than it in fact
is. T. at 354–55 (analogizing “the equivalent of too much clustering is putting dirt
on the lens of the telescope”). This Court finds Dr. Barber’s suggested level of
clustering had precisely this effect on Dr. Krosnick’s analysis; that is, in attempting
to control for the possible influence of factors not being tested, the clustering reduced
18
Despite Dr. Barber’s assertions that the .14 and .16 p-values render Dr. Krosnick’s
analysis unreliable, Dr. Barber admitted that he has reported coefficients as statistically significant
at a p-value of .15 in his own published work. T. at 733–36. This significantly impairs Dr.
Barber’s credibility on this point.
37
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 38 of 74
the accuracy of the investigation into the factors being tested for.19 This, Dr.
Krosnick explained, is known as “null hacking,” which “is hunting around for a way
to analyze the data to make an effect disappear,” such as by “clustering so severely
that the statistical power [of the data] is compromised in a way that an effect that’s
clearly present in the data is no longer statistically significant.” T. at 355. This is
precisely what Dr. Barber attempts to do here. His suggestions concerning clustering
are intended to muddy the water, not to clear it.
Dr. Krosnick’s analysis was further supported by testimony from Plaintiffs’
other experts. Plaintiffs also called Dr. Jonathan Rodden, a professor of political
science at Stanford University, as an expert in elections and data analysis. T. at 135–
40. Dr. Rodden testified concerning the effects of candidate name order on downballot races—both statewide and local races—where voters presumably have less
information about each candidate than they do about top-of-ticket candidates and
would, therefore, theoretically be more susceptible to candidate name order effects.
Specifically, Dr. Rodden compared the performance of down-ballot candidates with
top-of-ballot candidates of the same party to determine whether candidate name
order effects were more pronounced in either case. See Pls. Ex. 5 at 2–5. Dr. Rodden
19
As Dr. Barber acknowledged, this reduced accuracy could be linked to the relatively
small number of governor cycles (eleven since 1978) by which this analysis would be clustered.
T. at 736. It is therefore not the case that clustering this data would always reduce its accuracy,
but many more gubernatorial elections will have to occur before the full effect of such clustering
would be determinable with precision.
38
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 39 of 74
concluded down-ballot candidates who were listed second in their respective office
block garnered an average of 3.1% to 5.6% fewer percentage points of the vote than
did their top-of-ballot co-partisans in Florida elections. T. at 161; Pls. Ex. 5 at 4–5,
22. Dr. Rodden also found that, by contrast, down-ballot candidates who were listed
first in their respective office blocks consistently outperformed their top-of-ticket
co-partisans in Florida elections. T. at 152, 158–161, 170; Pls. Ex. 5 at 17–41. Dr.
Rodden testified he knew of no theory of political science which would explain these
results other than candidate name order effects. T. at 162–63. Dr. Barber also could
not provide any convincing explanation. See T. at 748–53. Dr. Rodden further
opined that, given the “heterogeneity” of the candidates in down-ballot races and the
various other factors that affect electoral outcomes, this result was “surprisingly
strong” and consistent across time and geography. T. at 175.
Dr. Barber attempted to call Dr. Rodden’s analysis into question by disputing
whether candidates’ personal qualities which might be difficult to quantify (such as
physical attractiveness, gender, regional origin, and roots in the community) played
a significant role in their electoral success, or lack thereof, and thus confounded Dr.
Rodden’s analysis. As proof of this, Dr. Barber pointed to the 2018 election for
Florida Commissioner of Agriculture, won by Democrat Nikki Fried. T. at 677; Ints.
Ex. 3 at 3–10.
39
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 40 of 74
Dr. Rodden, however, explained that he accounted in his analysis for the wide
variation in candidate qualities (such as height, physical attractiveness, and regional
origin) which might be less objectively measurable but which could still have been
reflected in the result. T. at 198–200. If this “heterogeneity” had a significant effect,
Dr. Rodden explained, one would expect it to “introduce a whole lot more noise and
make it even harder to find an effect.” T. at 200. But instead, the results showing
the impact of candidate name order effects were “surprisingly strong,” T. at 175,
meaning the statistical strength of the Dr. Rodden’s study was not affected by the
heterogeneity of the candidates in the races. Put more simply, the results of Dr.
Rodden’s study reflect the impact of candidate name order effects accurately and the
hard-to-quantify personal qualities of candidates did not meaningfully interfere with
his analysis. This also strengthen’s Dr. Rodden’s conclusions with respect to his
analysis of statewide races. As to the specific example of Commissioner Fried, this
Court finds it illogical to rely upon a single outlier to disprove a general conclusion,
particularly a conclusion about an average effect across a large number of samples.20
This Court finds Dr. Rodden’s methods and analysis reasonable, reliable, and
credible, and further finds Dr. Barber’s criticisms of Dr. Rodden’s analysis are
unreasonable and incredible.
20
Dr. Barber agrees that extrapolating from his example of Commissioner Fried would be
improper. T. at 759 (“Q. And you would agree that certainly you can’t draw conclusions about
trends among voters from one datapoint? A. I agree.”).
40
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 41 of 74
Dr. Rodden also investigated whether changing Florida’s ballot order scheme
might have a material impact on election results in Florida by comparing Florida to
North Carolina, a state which recently changed its ballot order scheme from a system
like Florida’s, where candidates of the governor’s party were listed first, to an
alphabetical system incorporating a random drawing to determine which letter to
begin the alphabetical ordering with. T. at 175–77; see also N.C. Gen. Stat. Ann.
§ 163A-1114(c) (2018) (requiring candidates to be listed in “either alphabetical
order or reverse alphabetical order by the last name of the candidate, which order
shall be determined each election by drawing at the State Board of Elections and
Ethics Enforcement after the closing of the filing period for all offices on the ballot”).
This resulted in a roughly even split between Republican and Democratic candidates
being listed first in North Carolina’s voting precincts. T. at 178. Comparing the
results of precincts in 2016, when Republican candidates were listed first in all
partisan races, with the results from those same precincts in 2018, when Republicans
were listed first only in roughly half of the precincts, Dr. Rodden found Democrats
obtained approximately 1.5% more of the votes cast in those precincts in which
Republicans were no longer listed first in 2018, than they had in 2016 when
Republicans were listed first in those precincts. Pls. Ex. 5 at 6, 45–46. In seats
where the incumbent was not running, this effect was as large as eight percentage
points, and in races where the same candidates faced each other in 2018 that had
41
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 42 of 74
faced each other in 2016 the effect averaged four percentage points. Id. at 6, 47–48.
Dr. Rodden testified that these results
gave [him] a clear sense of causality associated with reform and ballot
order practices. It allowed [him] to see what happens, holding a lot of
other things constant . . . when we just compare the change in the
Democratic vote share from 2016 to 2018; we can compare that in the
so-called treatment and control precincts and get a very clean causal
inference about the impact of reform.
T. at 179. Dr. Barber not only did not offer any substantive critique of this analysis,
he has also relied upon comparisons between Florida and North Carolina in the
course of his expert testimony in a North Carolina case. T. at 706–08. This Court
finds Dr. Rodden’s methods and conclusions to be reasonable, reliable, and credible.
This Court also finds that Dr. Barber’s testimony did not in any way undercut Dr.
Rodden’s analysis, and that Dr. Rodden’s analysis buttresses Dr. Krosnick’s analysis
by reinforcing the conclusion that candidate name order effects award a statistically
significant electoral advantage to the first-listed candidate, and that these effects are
even more pronounced in down-ballot races.
Finally, Plaintiffs called Dr. Paul Herrnson, a professor of political science at
the University of Connecticut, to explain an additional aspect of the advantage
candidate name order effects award to first-listed candidates. T. at 405–11; Pls. Exs.
4–8. Dr. Hernnson testified that voters intending to vote for the first-listed candidate
make fewer “proximity errors” than do voters intending to vote for candidates listed
in lower positions within an office block. T. at 418–21. Proximity error is a simple
42
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 43 of 74
human error in filling out a ballot which “results in the voter being unable to
accurately translate his or her voting intention into the candidate or, in some cases,
ballot issue they hope to select.” T. at 413. Instead, they unintentionally select the
option immediately before or immediately after the option they intend to select. Id.
Dr. Herrnson testified that this sort of error primarily advantages candidates listed
first in each office block and disadvantages those listed in subsequent positions. Id.
In races with more than two candidates, the second-listed candidate is most
disadvantaged. T. at 419. The reason for this is simple. Voters will make a number
of proximity errors due to factors such as fatigue, haste, or inattention, but if a voter
intends to vote for the first- or last-listed candidate, they can only err in one direction,
while voters intending to select the second-listed candidate out of three can err in
either direction. See T. at 417–20. Dr. Herrnson also explained the empirical basis
for this effect, including an experiment he himself conducted to investigate its
effects. T. at 420–35.
Dr. Barber opined that Dr. Herrnson did not address candidate name order
effects of the type at issue in this case. T. at 698. This Court disagrees. Dr.
Herrnson’s study of proximity error helps provide an empirical explanation for a
portion of the advantage a first-listed candidate receives by virtue of being listed first
in their office block. One reason candidate name order effects favor first-listed
candidates is that, in certain elections, voters are more likely to make errors in favor
43
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 44 of 74
of first-listed candidates and are also less likely to make errors which disadvantage
first-listed candidates.
In this way, the systematic placement of one party’s
candidates in the second position within their respective office blocks forces those
candidates to labor under the double disadvantage of the primacy effect and of a
greater likelihood of proximity error to their detriment. The effect of Florida’s ballot
order scheme, then, is to systematically place the candidates affiliated with the party
of the last-elected governor on the favorable end of these errors, solely on the basis
of their political affiliation. This Court finds Dr. Herrnson’s testimony credible, and
his methods and conclusions reasonable and reliable. This Court further finds Dr.
Herrnson’s testimony gives further support to the credibility of Dr. Krosnick’s
analysis by reinforcing the conclusion that candidate name order effects influence
the share of the vote a candidate receives, and particularly that primacy effects award
a statistically significant electoral advantage to the first-listed candidate.
After a full and careful review of the evidence in the record in this case, this
Court finds Dr. Krosnick’s testimony credible and his methods and conclusions
reasonable and reliable. Dr. Krosnick’s conclusions are supported by, and consistent
with, virtually all the decades of academic literature in the record, which is itself
almost universally consistent.
Furthermore, this Court finds Dr. Krosnick’s
evaluations of his original analysis, into which he incorporated Dr. Barber’s
concerns, to be reasonable, reliable, and credible. This Court also finds Drs. Rodden
44
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 45 of 74
and Herrnson’s testimony credible, and their methods and conclusions reasonable
and reliable. This Court further finds Dr. Barber’s testimony emphatically not
credible and his opinions offered in this case to be unreliable. This Court finds that,
even when Dr. Krosnick indulged Dr. Barber’s speculative critiques of his initial
analysis, Dr. Krosnick’s results did not undergo a material change, which further
supports Dr. Krosnick’s credibility, as does the fact that Dr. Krosnick performed that
second analysis to investigate the validity of Dr. Barber’s suggestions. This Court
further finds Dr. Barber’s criticisms of Dr. Rodden’s testimony to be disproven to
the extent they were not wholly speculative, and illogical in their reliance on a single
outlier to disprove a general rule. Additionally, this Court finds Dr. Barber’s opinion
concerning the significance of Dr. Herrnson’s analysis to be unpersuasive.
Plaintiffs have proven—and this Court hereby finds—that candidates of the
major parties in Florida receive an average primacy effect vote of approximately five
percent when listed first in their office block on the ballot, and that this advantage
accrues to a candidate because of the candidates’ name order, which in turn is
prescribed by section 101.151(3)(a), Florida Statutes. Furthermore, given Florida’s
history of election results in which the margin of victory or defeat is less than three
to five percentage points, this Court finds section 101.151(3)(a) has impacted
Plaintiffs’ First and Fourteenth Amendment rights by systematically allocating that
small but statistically significant advantage to Republican candidates in elections
45
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 46 of 74
where the last-elected governor was a Republican, just as it awarded that advantage
to Democrats in elections when Florida’s last-elected governor was a Democrat.
This Court need not find a precise percentage attributable to every election uniquely
to determine whether Florida’s ballot order scheme violates Plaintiffs’ rights,
particularly because Plaintiffs seek declaratory and injunctive relief. Rather, this
Court need only find—and does hereby find—that Florida’s ballot order statute
systematically awards a material advantage to candidates affiliated with the political
party of Florida’s last-elected governor solely on the basis of their party affiliation,
and therefore systematically disadvantages other candidates on the basis of their
party affiliation.
Additionally, the record reflects a vast number of Florida’s
elections have been decided by less than three to five percent of the votes cast—in
other words, by a smaller margin than the advantage Florida’s ballot order scheme
awards to the candidates affiliated with the party of Florida’s last-elected governor.
Defendants offer no reason to expect the future will be any different, or that
Floridians are any different from voters in other states or countries (with the apparent
exception of Afghanistan) with respect to their susceptibility to psychological
phenomena like candidate name order effects.
In summary, this Court finds Plaintiffs have proven the candidate listed first
in their respective office block in Florida elections receives, on average, a fivepercentage-point advantage over their competitors for that office by virtue of being
46
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 47 of 74
the first-listed candidate. This Court further finds that, even when Dr. Barber’s
speculative critiques are incorporated, first-listed candidates receive an advantage of
approximately three-and-a-half percent. This Court finds these amounts to be
statistically significant and the calculations which produced them to be reliable,
reasonable, and credible. This Court finds Dr. Krosnick, Dr. Rodden, and Dr.
Herrnson credible, and finds their respective analyses reinforce and add depth to one
another’s conclusions. This Court finds Dr. Barber’s critiques do not undermine
Plaintiffs’ experts’ methods, analyses, or conclusions. This Court finds Dr. Barber’s
critiques speculative, unreasonable, and not credible.
Although these findings are based exclusively on the evidence in the record
before this Court, it is worth noting that they are hardly an aberration. Other courts
confronted with this question have also discussed the influence of ballot position on
candidates’ electoral outcomes and reached similar conclusions. See, e.g., McLain
v. Meier, 637 F.2d 1159, 1166 (8th Cir. 1980) (stating “[i]n concluding that there
was no error in the district court’s finding of ballot advantage in the first position we
are not the first to so hold,” and citing numerous cases). And a majority of the States
use nonpartisan ballot order schemes that would tend to minimize the impact of
candidate name order effects, presumably to control for this widely understood
47
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 48 of 74
phenomenon.21 There is nothing new or surprising about the conclusion that
Florida’s ballot order scheme confers a small but statistically significant advantage
to candidates of one political party. Finally, as previously noted, Intervenors clearly
believe some advantage presently accrues to Republican candidates as a result of
Florida’s ballot order scheme. Otherwise, they would not have intervened in this
case and would not be spending such significant time and resources defending the
statute. This fact casts doubt on Intervenors’ assertion that Florida’s ballot order
scheme does not have any significant effects and changing Florida’s ballot order
scheme would not affect Republican electoral efforts. It is also entirely consistent
with Dr. Krosnick and Dr. Rodden’s conclusions—which this Court finds credible—
that Florida’s ballot order does indeed make a difference to the outcome of elections
in Florida.
The advantage conferred by the primacy effect is relatively small percentage
taken in isolation, but the records of Florida’s elections which are before this Court
demonstrate it is more than the margin of victory or defeat in a great many elections
21
See Ala. Code § 17-6-25; Alaska Stat. § 15.15.030(6); Ark. Code Ann. § 7-5-207(c)(1);
Cal. Elec. Code § 13111; Colo. Rev. Stat. § 1-5-404; Haw. Rev. Stat. § 11-115; Idaho Code Ann.
§§ 34-903(4) & 34-2419; Kan. Stat. Ann. § 25-610; Ky. Rev. Stat. Ann. § 118.225; La. Rev. Stat.
Ann. § 18:551(C); Me. Rev. Stat. Ann. tit. 21-A, § 601; Miss. Code Ann. § 23-15-367(2); Mont.
Code Ann. § 13-12-205(2); Nev. Rev. Stat. § 293.267; N.H. Rev. Stat. Ann. § 656:5(II); N.J. Stat.
Ann. § 19:14-12; N.M. Stat. § 1-10-8.1; N.C. Gen. Stat. Ann. §163A-1114; N.D. Cent. Code
§§ 16.1-11-27 & 16.1-06-05; Ohio Rev. Code Ann. § 3505.03; Okla. Stat. Ann. tit. 26, § 6-106;
Or. Rev. Stat. § 254.155; R.I. Gen. Laws § 17-19-9.1; S.D. Codified Laws § 12-16-3.1; Utah Code
Ann. § 20A-6-302(1)(b); Vt. Stat. Ann. tit. 17, § 2472; Va. Code Ann. § 24.2-613.
48
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 49 of 74
in Florida. Of course, many factors determine the total percentage of the vote a
candidate ultimately receives, but Plaintiffs’ experts have proven to a high degree of
probability that a small but statistically significant portion of the votes earned in any
election are attributable to the primacy effect. It would be unreasonable to contend
that candidate name order effects are the only reason a candidate wins or loses an
election. There are a multitude of factors which contribute to a candidate’s success,
or lack thereof, but as Dr. Krosnick explained, candidate name order effects have
often been outcome determinative. See T. at 358–63. But although no single
raindrop bursts a dam, and a single small transaction rarely is the sole cause of a
bankruptcy, the dam still fails and the debtor still becomes insolvent. Similarly,
candidate name order effects are not the only reason elections are won and lost, but
they do contribute substantially to candidates’ successes or failures at the polls.
As to the character of the injury, this Court concludes section 101.151(3)(a)
is discriminatory because it awards the primacy effect vote to candidates based
solely and uniquely upon their political affiliation.
By listing the candidates
affiliated with the governor’s party first in each partisan race, Florida’s current ballot
order scheme ensures those candidates receive a statistically significant advantage
in that election, conferred by the primacy effect; and it does so on the explicit basis
of the candidates’ party affiliation. It is in no way politically neutral. Intervenors
argue the statute is nonpartisan because it can benefit—and historically has
49
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 50 of 74
benefitted—whichever party controls the Governor’s Mansion. But this does not
make the statute nonpartisan.
Rather, it means the direction of the statute’s
partisanship changes depending on circumstances.22 Florida’s ballot order scheme
is a fair-weather friend to the party in power, whichever party that may be, and
although its inclination may change depending on the prevailing political breeze it
is unquestionably a partisan provision. The character of Plaintiffs’ injury, therefore,
is politically discriminatory. By systematically awarding a statistically significant
advantage to the candidates of the party in power, Florida’s ballot order scheme takes
a side in partisan elections.
Legitimacy and Strength of Justifications
Next, this Court must “identify and evaluate the precise interests put forward
by the State as justifications for the burden imposed” by section 101.151(3)(a).
Anderson, 460 U.S. at 789. This Court must consider whether those justifications
are legitimate, determine how strong the interests they support are, and examine the
extent to which those interests make it necessary to impose these specific burdens
on Plaintiffs’ rights. Id.
Defendant Lee asserts Florida’s present ballot order scheme is supported by
Florida’s interest in “upholding the policy choices of Florida’s duly-elected
22
By Defendants’ logic, because he held opposite allegiances at different times, Benedict
Arnold was a neutral party.
50
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 51 of 74
representatives,” preventing voter confusion, “promoting uniformity,” and
“promoting voter confidence in the integrity of the election administrations process.”
ECF No. 199 at 34. 23 Some of these overlap. For example, Defendant Lee argues
Florida’s current ballot order scheme promotes voter confidence by ensuring “people
know that their ballot is being arranged consistent with the choices their elected
officials made, in a manner that makes it easy to find candidates of their choice on
the ballot, in a manner that is uniform throughout the State, and in a manner that
allows for accurate vote tabulation.” Id. Similarly, Florida’s asserted interest in
uniformity is justified in part by its tendency to “reduce voter confusion.” Id.
It is important to note at the outset that Defendant Lee’s asserted justifications
for Florida’s current ballot scheme are not quite on point with the Anderson/Burdick
standard. That framework requires this Court to consider whether the state’s
asserted interests “justify the specific restriction . . . at issue.” Anderson, 460 U.S.
at 796 (emphasis added). For the most part, Defendant Lee asserts justifications
which would tend to support any ballot order system—or, more generally, any
election law at all—and militate against changing it to another; but they are not, for
the most part, justifications for the specific ballot order scheme Florida uses as
opposed to any alternative one. In effect, Defendant Lee asks this Court to weigh
23
In their closing argument, Intervenors relied in large part on Defendant Lee’s arguments
with respect to the justifications for Florida’s current ballot order scheme. See ECF No. 200 at 27.
51
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 52 of 74
the burden on Plaintiffs’ rights against Florida’s general interest in having election
laws, whatever they may be, without reference to the unique necessity for those laws.
What Anderson/Burdick requires this Court to do is to compare the specific burden
imposed by a challenged law to its specific justifications and necessity, and then
determine which is weightier (with, of course, the appropriate level of scrutiny
calibrating the scale).
Take, for example, the asserted interest in “promoting uniformity to reduce
errors in ballot layout and vote tabulation, and to reduce voter confusion.” ECF No.
199 at 34. To the extent Plaintiffs suggest candidates’ names should be rotated
within office blocks to reduce any overall primacy effect, this could be interpreted
as reducing uniformity; but under that system Florida’s ballots would be organized
according to a single framework, and notwithstanding any rotation, that would
nevertheless be a uniform system of ballot order. Even more obviously, the asserted
interest in promoting voter confidence is, in plain English, an interest in ensuring
that voters know their ballot is organized as the law requires it to be. This has
nothing to do with the substance of how the law actually requires the ballot to be
organized, so long as it is organized according to some system imposed by law. It
does not militate in favor of any particular system of organization. Therefore, to the
extent they are true of any ballot order scheme and not specifically related to the
52
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 53 of 74
scheme laid out in section 101.151(3)(a), Florida Statutes, this Court finds Defendant
Lee’s asserted justifications to be extremely weak.
To the extent they relate specifically to Florida’s current ballot order scheme,
the justifications Defendant Lee asserts are a mixed bag. First, the interest in
“upholding the policy choices of Florida’s duly-elected representatives,” id., is not
particularly persuasive. Florida certainly has a valid and important sovereign
interest in defending the Florida Legislature’s enactments, but the undoubted
importance of this interest should not be conflated with the weight it should be
accorded under the Anderson/Burdick standard. If the Florida Legislature adopts an
unconstitutional law, Florida’s interest in having made that legislative choice will
not render that law constitutional. See Marbury, 1 Cranch at 177–80 (explaining the
foundations of judicial review of legislative action); Nat’l Foreign Trade Council v.
Natsios, 181 F.3d 38, 61 (1st Cir. 1999) (“Even where they exist, strong state
interests do not make an otherwise unconstitutional law constitutional.”). A state’s
legitimate interests may prevent a law which impacts constitutional rights from
rising to the level of a constitutional violation, see, e.g., Anderson, 460 U.S. at 789
(requiring courts to balance the burdens on individuals’ rights against state’s
interests and justifications to determine whether the challenged law violates the
Constitution); but a state has no legitimate interest in preserving and enforcing a law
53
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 54 of 74
which violates the Constitution.24 See Hunter v. Hamilton Cty. Bd. of Elections, 635
F.3d 219, 238 (6th Cir. 2011) (concluding disparate treatment of voters resulted “not
from a ‘narrowly drawn state interest of compelling importance,’ but instead from
local misapplication of state law” (quoting Crawford, 553 U.S. at 190)); Bishop v.
Holder, 962 F. Supp. 2d 1252, 1289–90 (N.D. Okla. 2014) (concluding that,
although a provision of the Oklahoma Constitution which discriminated against
homosexuals “rationally promoted the state’s interest in upholding one particular
moral definition of marriage, this is not a permissible justification” for the provision
under rational-basis scrutiny). Defendant Lee essentially argues Florida’s current
ballot order scheme is necessary because the Florida Legislature has decided it is.
Although legislative discretion to formulate policy is an important element
underpinning this analysis, it is not dispositive of the outcome, particularly in cases
such as the present one where the burden on voting rights is not minimal. See
Burdick, 504 U.S. at 434. This Court does not discount this justification entirely,
but accords it little weight in view of the fact it bears only tangentially upon the
merits of the issue.
24
Defendant Lee does not argue that Florida’s ballot order statute reduces confusion by
making it more convenient for Republicans to locate their preferred candidates, and Florida would
have no legitimate interest in such a goal. See McLain, 637 F.2d at 1167 (explaining that the
asserted interest in placing incumbents first “virtually admits that the state has chosen to serve the
convenience of those voters who support incumbent and major party candidates at the expense of
other voters”).
54
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 55 of 74
Second, Florida’s asserted interest in preventing voter confusion is legitimate,
but not particularly strong on the specific facts of this case. The only evidence
Defendants offered in favor of Florida’s current ballot order statute was speculation
that, in a county-by-county rotation scheme, some voters might be confused by
seeing sample ballots from other counties. This Court finds those suggestions are
not credible. Sample ballots state which county they pertain to, so even in areas with
multicounty media markets, voters are unlikely to become confused and are even
less likely to be more confused by county-by-county rotation than they are by the
differences between counties’ ballots in the current system. And even within a
county, ballots vary from precinct to precinct, yet voters do not find it confusing that
their precinct’s ballot does not match exactly the county-wide sample ballot
provided by their respective supervisor of elections. Finally, any minimal concerns
about increased voter confusion under a county-by-county or precinct-by-precinct
rotational system are entirely nullified in a system which arranges candidates in a
uniform manner statewide according to alphabetical order within office blocks.25
This Court finds this justification relevant, but concludes it should be accorded only
minimal weight at most.
25
As a logical matter, the same would be true of systems which order candidates’ names
in order of qualification or by random lottery.
55
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 56 of 74
Florida’s asserted interest in promoting uniformity is, though legitimate,
similarly weak here. At the heart of this asserted interest is the desire to minimize
tabulation errors and avoid the costs associated with modifying Florida’s election
software and hardware to facilitate different ballot order schemes. The upshot of the
testimony presented in this case is, although it may be possible for Florida to adapt
its election machinery to rotate candidates’ names from county to county or from
precinct to precinct, no one is sure precisely what that would entail or what hidden
gremlins might emerge in the process. It is apparent that Florida’s voting systems
would have to be recertified were such a rotational scheme to be employed, and
witnesses raised further concerns regarding the possible impacts on the amount of
time necessary to prepare ballots in individual counties. It is not, however, certain
that rotation would increase the danger of tabulation errors, since votes are tabulated
based on a unique candidate number assigned at the state level and not by the county,
but it was unclear what precise adjustments, if any, would be necessary to Florida’s
tabulation mechanisms to permit county-by-county or precinct-by-precinct rotation
to be used. This does not, however, translate into a necessity for Florida’s present
ballot order scheme. Rather, it is evidence of the predicted administrative burden
associated with changing that entrenched ballot order scheme to one possible
alternative.
56
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 57 of 74
This is not the perspective the Anderson/Burdick analysis requires this Court
to take. Instead, Anderson/Burdick requires courts to consider the necessity for, and
justifications supporting, a particular voting regulation on its own merits, not merely
because it is the status quo. That said, to the extent it relates to Florida’s present
ballot order scheme per se, this justification demonstrates that a county-by-county
or precinct-by-precinct rotational system would impose administrative burdens that
Florida’s current ballot order scheme avoids. This weighs in favor of Florida’s
current ballot order scheme as compared with those alternatives. With respect to
alphabetical ordering, ordering by date and time of qualification, and order based on
random lottery, however, there is no similar concern about administrative burdens
or tabulation errors because candidates would be listed in a uniform order on all
ballots statewide.
It would not be necessary to adapt Florida’s election
administration machinery to adopt alphabetical ordering, because Florida already
uses it in certain elections. See § 101.151(4)(a), Fla. Stat. (2019).26 Therefore,
although Florida’s legitimate interest in uniformity carries moderate weight with
regard to county-by-county and precinct-by-precinct rotation, it carries none with
26
The same is true for order-of-qualification ballot ordering. Id. § 101.151(3)(b).
Although it was not specifically dealt with in depth during the trial of this cause, there is no reason
for this Court to suspect ordering candidates by date and time of qualification or by random lottery
would present any more difficulty than the current system, either.
57
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 58 of 74
respect to nonpartisan alternatives such as alphabetical listing of candidates within
office blocks.
Florida’s interest in uniformity also relates to testimony elicited during trial
that employing county-by-county rotation would expose both the State and the
individual county supervisors of elections to increased litigation by candidates
dissatisfied with their ballot placement. See T. at 464–68. Defendants’ concern is
that candidates for offices whose districts are entirely contained within a single
county would not benefit from any rotation of names, and would therefore
potentially sue both the State and their local supervisor of elections. This is no
justification at all for Florida’s current ballot order scheme. First, any change to an
existing statute will produce litigation concerning the meaning of the new version,
and this justification therefore does not relate to Florida’s specific ballot order
scheme as opposed to any alternative one. This Court doubts any law, no matter
how carefully drafted, will ever be so clear that a sufficiently clever attorney will
find nothing in it to be litigated. Second, even assuming it is relevant and legitimate,
this justification is particularly weak because, on its face, county-by-county rotation
could only reduce the amount of potential litigation related to Florida’s ballot order
scheme. Under Florida’s current ballot order scheme, any second-listed candidate
in any partisan race could sue both the Secretary of State and each county supervisor
of elections on whose ballots they appear. In a county-by-county rotational system,
58
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 59 of 74
meanwhile, only second-listed candidates contesting elections in districts
encompassing a single county or part of a single county—that is, districts whose
candidates’ names would not be rotated—would have a claim to bring. It is plain
that county-by-county rotation would in fact reduce the possibility of litigation by
significantly reducing the number of possible plaintiffs. This Court finds the witness
testimony to the contrary to be oppugnant to common sense, and not credible. It is
equally plain that, for the same reasons, an alphabetical system of candidate name
order would reduce the possibility of litigation still further. This justification is too
weak to bear even the mild scrutiny this Court employs in this Order.
Finally, a state has no legitimate interest whatsoever in promoting its citizens’
confidence in adherence to an unconstitutional law; but to the extent Florida’s
asserted interest in this case is legitimate and relevant, there is at best no reason to
believe Florida’s citizens draw more confidence from one sort of ballot order scheme
than from any other. If anything, voter confidence in the integrity of Florida’s
elections process would actually be undermined by Florida’s present ballot order
scheme. Taking Defendant Lee’s argument on this score at face value, Florida’s
ballot order would promote citizen confidence to the same degree whether it requires
candidates affiliated with the last-elected governor’s party to be listed first, as it does
now, or whether it lists them alphabetically, in order of qualification, or completely
randomly. Although some testimony was presented to suggest a county-by-county
59
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 60 of 74
rotation scheme would reduce voter confidence in the election system, this Court
finds that testimony not credible for two reasons. First, Defendant Lee links voter
confidence to voters knowing that the ballot is ordered as the law requires.
Therefore, even if the candidates are rotated on a county-by-county basis, voters
presumably are no less ignorant from one day to the next and will gain confidence
from knowing the ballot was organized as the law requires. Second, this Court
concludes voters could actually lose confidence in the integrity of Florida’s election
system from the knowledge that the order of candidates’ names on the ballot was
determined on a partisan political basis rather than a neutral one. With respect to the
necessity for Florida’s present ballot order scheme in terms of voter confidence,
therefore, this Court finds that, at best, Plaintiffs have shown there is no necessity
for it vis-à-vis the alternatives, and at worst Florida’s present ballot order scheme is
less effective at promoting voter confidence than several alternatives.
Scrutinizing the Statute
Taking all this together, this Court must now determine which level of
scrutiny to apply and balance the burdens on Plaintiffs’ rights against Defendants’
asserted justifications for those burdens. In the Anderson/Burdick framework, when
a court concludes a law imposes only “reasonable, nondiscriminatory restrictions”
on voting rights, Anderson, 460 U.S. at 789, the court scrutinizes the challenged law
only on a level akin to rational basis review. Id. If, however, the burden is more
60
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 61 of 74
severe, the level of scrutiny undergoes a concomitant increase. Burdick, 504 U.S. at
434. The Anderson/Burdick framework thus incorporates a sliding scale of scrutiny
linked expressly to the burden on the plaintiff’s rights.
As burdens on voting rights go, Florida’s ballot order scheme is not on the
same order of magnitude as a complete denial of the franchise, such as preventing
someone from registering to vote or from casting their vote. As Defendants correctly
contend, it does not prevent any individual from voting, nor does it prevent their
votes from being counted. Under similar circumstances, other courts have found it
appropriate to apply a rational-basis standard. See McLane, 637 F.3d at 1167
(reasoning that “although ballot format has an effect on the fundamental right to
vote, the effect is somewhat attenuated” and explaining that “[i]n these
circumstances, most courts have applied the rational basis test”). The present case
is different for two reasons. First, as explained above, Florida’s ballot order statute
is not neutral; instead, it affects Plaintiffs’ rights in a politically discriminatory way.
Second, although a donkey vote of three or even five percent is not, in and of itself,
a large proportion of the total vote, it is often a decisive proportion in terms of the
spread between the candidates in a Florida election. This suggests that, although the
61
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 62 of 74
quantitative burden on Plaintiffs’ rights is small, the practical burden is severe
indeed. Cf. Anderson, 460 U.S. at 806 (conducting a “realistic” appraisal).27
This Court concludes the appropriate standard of review in this case is higher
than rational-basis review.
Florida’s ballot order scheme is not a neutral,
nondiscriminatory restriction on Plaintiffs’ voting rights, and the burden it imposes
is significant. But Florida’s ballot order scheme is not a well-nigh insuperable
burden, and Florida has legitimate regulatory interests in administering orderly
elections to which this Court owes respectful deference. This Court therefore
concludes neither mere rational-basis review nor strict scrutiny would be
appropriate. Instead, this Court will scrutinize the State’s asserted interests and
justifications in a level of detail greater than the cursory examination afforded under
the rational-basis standard, but this Court will not apply a stricter standard that
requires the State’s interest to be compelling or important—they need only be
legitimate and relevant; and (pursuant to the Anderson/Burdick framework’s inquiry
into the necessity for the challenged restriction) the challenged law must actually
serve the State’s asserted interest to at least the same degree as an available
alternative, otherwise it would not be necessary. Nor will this Court interpret the
27
In this same sense, a $1,000 medical bill might seem insignificant or insuperable
depending on one’s monthly income. Severity is an objective question but the realistic context is
important to the analysis.
62
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 63 of 74
“necessity” aspect of the Anderson/Burdick framework to signify an absolute
necessity in this context, as stricter shades of scrutiny would require. Finally, this
Court will not pry into the subjective motivations the State of Florida had in adopting
its current ballot order scheme, but will instead consider only the objective interests
and justifications Defendants raise and evaluate whether they are relevant,
legitimate, and of sufficient strength to justify the specific burden at issue. Under
the Anderson/Burdick framework, this is the correct analysis for this Court to apply.
Applying that level of scrutiny, this Court concludes Florida’s interests in its current
ballot order scheme do not justify the burden on Plaintiffs’ rights which Florida’s
current ballot order scheme imposes. This Court also concludes, however, that
Florida’s current ballot order scheme cannot satisfy even the extremely deferential
rational-basis standard under the Anderson/Burdick framework.
Put simply,
Florida’s current ballot order scheme cannot even clear the lowest available bar.
As this Court explained above, Florida’s current ballot order scheme imposes
a burden on Plaintiffs’ First and Fourteenth Amendment rights which, although
numerically small, is significant in both the statistical sense and in qualitative terms.
Furthermore, that burden is of a politically discriminatory character. To weigh
against that burden, Defendants have offered a variety of justifications which this
Court has found vary in strength from minor to meager, to the extent they are
relevant and legitimate at all. Weighing the burden on Plaintiffs’ rights against the
63
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 64 of 74
strength (such as it is) of Florida’s interests in, and the necessity of, imposing that
burden, this Court concludes Florida has advanced no sufficient justification nor
combination of justifications nor a necessity, reasonable or otherwise, for employing
a ballot order scheme which systematically advantages candidates of one party and
disadvantages candidates of another party to a statistically significant (and
potentially outcome-determinative) degree. In light of the almost total absence of
any legitimate state interest favoring this system over the multiple alternatives
available, it is difficult to imagine what other purpose it could possibly serve than as
a thumb on the scale in favor of the party in power. This Court holds the burden on
Plaintiffs’ First and Fourteenth Amendment rights conclusively outweighs
Defendants’ asserted interests and justifications, which are universally weak.
Moreover, even under the rational-basis standard, Florida has not advanced relevant,
legitimate interests which the Florida Legislature could rationally conclude justify
burdening Plaintiffs’ rights as Florida’s current ballot order scheme does. Under
either standard, Florida’s current ballot order scheme violates the First and
Fourteenth Amendments.
Permanent Injunction
This Court has concluded Florida’s current ballot order scheme is
unconstitutional. The question now is, what to do about it? Plaintiffs seek a
permanent injunction to prohibit Defendant Lee from enforcing section
64
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 65 of 74
101.151(3)(a) as it currently stands.
Plaintiffs also ask this Court to require
Defendant Lee to adopt a ballot order scheme of county-by-county rotation of
candidates’ names until such time as the State of Florida adopts a replacement ballot
order scheme. To obtain a permanent injunction, Plaintiffs must show that (1) they
have suffered an irreparable injury; (2) their remedies at law are inadequate; (3) the
balance of hardships weighs in their favor; and (4) that a permanent injunction would
not disserve the public interest. eBay, Inc. v. MercExchange, L.L.C., 547 U.S. 388,
391 (2006). Plaintiffs have satisfied each element of this standard.
“[T]he irreparable-injury requirement may be satisfied by demonstrating a
history of past misconduct, which gives rise to an inference that future injury is
imminent.” Thomas v. Bryant, 614 F.3d 1288, 1318 (11th Cir. 2010). Moreover,
“the irreparable-injury requirement cannot be met absent a real or immediate threat
that the plaintiff will be wronged again.” Id. Injuries are irreparable if they cannot
be undone through monetary remedies. Cunningham v. Adams, 808 F.2d 815, 821
(11th Cir. 1987). And irreparable injury is presumed when “[a] restriction on the
fundamental right to vote” is at issue. Obama for Am. v. Husted, 697 F.3d 423, 436
(6th Cir. 2012). This Court has elsewhere explained that injuries to fundamental
voting rights are irreparable because “[t]his isn’t golf: there are no mulligans.” Fla.
Democratic Party II, 215 F. Supp. 3d at 1258. This Court finds Plaintiffs have
conclusively demonstrated a history of injury and have established a real and
65
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 66 of 74
immediate threat that, absent equitable relief from this Court, they will be wronged
again. Plaintiffs satisfy the first factor.
Second, remedies at law are not sufficient to redress Plaintiffs’ injuries.
Damages would in no way remedy these violations of Plaintiffs’ rights.
Cf.
Cunningham, 808 F.2d at 821. The injury Plaintiffs suffer is the result of holding
elections with ballots organized pursuant to Florida’s current ballot order scheme.
This Court finds Plaintiffs have satisfied the second factor.
Third, the balance of hardships weighs in Plaintiffs’ favor. As this Court has
explained at length, Plaintiffs suffer from a substantial burden on their First and
Fourteenth Amendment rights, and the countervailing hardship the State of Florida
would suffer ranges from minimal to nonexistent. See Laube v. Haley, 234 F. Supp.
2d 1227, 1252 (M.D. Ala. 2002) (“The threat of harm to the plaintiffs cannot be
outweighed by the risk of financial burden or administrative inconvenience to the
defendants.”). This Court finds Plaintiffs have satisfied the third factor.
Fourth and finally, a permanent injunction will not disserve the public interest.
It is beyond question that the public’s interest is served by upholding constitutional
rights. Cf. Laube, 234 F. Supp. 2d at 1252 (“[T]here is a strong public interest in
requiring that the plaintiffs’ constitutional rights no longer be violated . . ..”);
Republican Party of Minn., 416 F.3d at 753 (“It can hardly be argued that seeking to
uphold a constitutional protection . . . is not per se a compelling state interest.”); see
66
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 67 of 74
also League of Women Voters of Fla. v. Browning, 863 F. Supp. 2d 1155, 1167 (N.D.
Fla. 2012) (“The vindication of constitutional rights . . . serve[s] the public interest
almost by definition.”).
Voting is at the heart of representative systems of
government, and dispensing with unconstitutional obstacles to free and fair elections
is unquestionably in the public interest. Cf. Fla. Democratic Party II, 215 F. Supp.
3d at 1258 (quoting Reynolds v. Sims, 377 U.S. 533, 555 (1964)). This Court finds
Plaintiffs have satisfied the fourth factor.
This Court concludes Plaintiffs are entitled to a permanent injunction, and
must next consider what the terms of that injunction will be. Courts have broad
discretion to fashion equitable relief which is appropriate to the circumstances of the
case in question. See Democratic Exec. Comm., 915 F.3d at 1327–28 (explaining
that crafting a “Goldilocks solution,” rather than approaching equitable relief as an
all-or-nothing proposition, is “well within [a court’s] discretion”).
The
circumstances of the present case make this a relatively simple process. Plaintiffs
seek a permanent injunction which both prohibits enforcement of Florida’s current
ballot order scheme and requires Defendant Lee to employ a county-by-county
rotational scheme until such time as the State of Florida adopts a new ballot order
scheme on a permanent basis. See ECF No. 201-1 at 62–63. This Court concludes
a permanent injunction against enforcement or application of section 101.151(3)(a)
is warranted, but that an injunction requiring Defendant Lee to specifically adopt
67
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 68 of 74
county-by-county rotation until the State of Florida chooses a replacement would
not be appropriate.
The alternatives for ballot order schemes can be divided into two categories
based upon the way in which they would reduce the burden on, or altogether avoid
burdening, Plaintiffs’ rights, while also serving Florida’s legitimate regulatory
interests.
The first category are the rotational schemes, which would rotate
candidates’ names within their office blocks on a county-by-county or precinct-byprecinct basis. These schemes attempt to minimize the burden on Plaintiffs’ rights
by distributing the candidate name order effects more evenly across all candidates
rather than awarding it solely to candidates of one party. Rotational ballot order
schemes satisfy the requirements of the First and Fourteenth Amendment by
equalizing the burden on voting rights—they do not remove it, but they do burden
all persons equally to the extent reasonably possible. This is one set of ways to solve
the problem.
The second category of ballot order schemes alleviate the burden on First and
Fourteenth Amendment rights by cleansing the partisan taint from the process. If
candidates’ names are arranged in alphabetical order, or in the order in which
candidates submitted their qualifying paperwork, or in an order determined by
random lottery, the “donkey vote” would still exist and its magnitude would be
unchanged; but it would not be distributed on the basis of candidates’ political
68
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 69 of 74
affiliation, and no constitutional issue would arise on that basis. With respect to this
category of ballot order schemes, the balance is most favorable to Plaintiffs because
Defendants’ justifications are at their weakest.
This Court will not require that Florida adopt a particular ballot order scheme
from among the alternatives described above, nor that Florida choose only from
among the alternatives this Court and the parties have identified. If there were only
one alternative ballot order scheme which satisfied the Constitution, this Court’s
conclusion on this question might be different; but, as described above, there are
several alternatives which do not systematically award the advantage of the primacy
effect vote to candidates of one party based on their party affiliation and which do
not impose significant burdens on the State. The State of Florida is empowered, by
its sovereignty, to adopt any ballot order scheme that comports with the requirements
of the Constitution and other applicable law. Cf. Democratic Exec. Comm., 915 F.3d
at 1331 (explaining that, “rather than undermining Florida’s sovereignty,” the
narrowly tailored injunction issued in that case “actually respected it”). Moreover,
the issue in this case is not whether Florida is failing to do something it ought to do,
but rather whether the Constitution permits it to act in a particular way. Therefore,
this Court concludes a negative injunction rather than an affirmative one is
appropriate.
69
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 70 of 74
Florida may choose to adopt county-by-county rotation, precinct-by-precinct
rotation, alphabetical-by-surname listing, or any number of other alternative ballot
order schemes which comport with the Constitution. But Florida may not continue
to order candidates’ names on its ballots in a way that systematically awards the
statistically significant advantage conferred by candidate name order effects to
candidates of one political party on the basis of their partisan affiliation. 28 As this
Court has previously noted, “[w]hile the vast majority of supervisors of election are
upstanding professionals who follow the law and court orders . . . there may be some
who selectively interpret parts of this Court’s orders or otherwise avoid compliance.”
Rivera Madera v. Lee, No. 1:18cv152-MW/GRJ, 2019 WL 2077037, at *3 (N.D.
Fla. May 10, 2019). This Court will therefore be blunt. Compliance with this Order
is not optional, and this Court will not hesitate to use every tool available to enforce
its authority and ensure compliance with its orders. See Fed. R. Civ. P. 65(d)(2)(C)
(binding “other persons who are in active concert or participation” with the parties
and their officers, agents, and employees); 28 U.S.C. § 1651(a) (2018) (endowing
courts with the authority to “issue all writs necessary or appropriate in aid of their
28
As to the immediate effect of the injunction this Court issues, concerns were expressed
during the trial of this case regarding whether Florida’s election systems could be recertified to
permit county-by-county or precinct-by-precinct rotation to be implemented in time for the 2020
general election. This Court notes that there are other options, such as alphabetical ordering or
ordering in order of qualification, that Florida already uses and therefore that Florida’s election
systems already are capable of, and thus those ballot order alternatives could be implemented
immediately. See §§ 101.151(3)(b) & 101.151(4)(a), Fla. Stat. (2019).
70
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 71 of 74
respective jurisdictions and agreeable to the usages and principles of law”). This
Court does not anticipate any such extraordinary measures will be necessary. It’s
not hard—the way for Florida to conduct a free and fair election is, to conduct a free
and fair election.29 Accordingly,
IT IS ORDERED:
1. The Clerk shall enter judgment stating: “This Court hereby DECLARES that
the ballot order scheme described in section 101.151(3)(a), Florida Statutes,
violates Plaintiffs’ rights under the First and Fourteenth Amendments of the
United States Constitution. This Court GRANTS Plaintiffs’ request for a
permanent injunction. Pursuant to the Secretary of State’s responsibility for
“general supervision and administration of the elections laws,” § 15.13, Fla.
Stat. (2019), and her authority to “[o]btain and maintain uniformity in the
interpretation and implementation of the elections laws,” § 97.012(1), Fla.
29
One could be forgiven for thinking otherwise, given the difficulty Florida seems to have
with the concept. See Rivera Madera, 325 F. Supp. 3d at 1283–84 (noting “[i]t is remarkable that
it takes a coalition of voting rights organizations and individuals to sue in federal court to seek
minimal compliance with the plain language of a venerable 53-year-old law,” the Voting Rights
Act); Fla. Democratic Party I, 2016 WL 6090943, at *1 (explaining some of the manifold ways
in which “the State of Florida has consistently chipped away at the right to vote”); Rivera Madera
v. Lee, 2019 WL 2077037, at *4 (explaining the Voting Rights Act “has been the law of the land
since 1965 and supervisors of elections should have been complying with the law for more than
50 years without court intervention.”); Hand, 285 F. Supp. 3d at 1310 (noting “[m]ore than one
tenth of Florida’s voting population—1.7 million as of 2016—cannot vote because they have been
decimated from the body politic” and “[m]ore than one in five of Florida’s African American
voting-age population cannot vote.” (internal marks and footnotes omitted)); Fla. Democratic
Party II, 215 F. Supp. 3d at 1257 (explaining Florida’s voter registration framework “completely
disenfranchise[d]” hundreds of thousands of potential voters by failing to include a provision
extending the registration period in the event of an emergency or natural disaster).
71
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 72 of 74
Stat. (2019), and pursuant to Federal Rule of Civil Procedure 65(d)(2)(C),
neither Defendant Lee, her successors in office, deputies, officers, employees,
agents, nor any other person in active participation or concert with Defendant
Lee shall enforce, nor permit enforcement of, the ballot order scheme
described in section 101.151(3)(a), Florida Statutes. Defendant Lee and her
successors in office, as well as her deputies, officers, employees, agents, and
any other person in active participation and concert with Defendant Lee shall
take all practicable measures within the scope of their official authority to
ensure compliance with the terms of this Order. From the date of this Order
forward, no ballot shall issue which is organized pursuant to the ballot order
scheme described in section 101.151(3)(a), Florida Statutes. No supervisor of
elections of any Florida county, nor their successors in office, deputies,
officers, employees, agents, nor designees, shall issue any ballot which is
organized pursuant to the ballot order scheme described in section
101.151(3)(a), Florida Statutes.”
2. Within fourteen days after the issuance of this Order, Defendant Lee shall
provide written guidance to the supervisors of elections of Florida’s counties
informing them that this Court has declared the ballot order scheme described
in section 101.151(3)(a), Florida Statutes, unconstitutional. Defendant Lee
72
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 73 of 74
shall include a true and correct copy of this Court’s order in her written
guidance.
3. Once Defendant Lee has provided said written guidance to the supervisors of
elections, Defendant Lee shall file a notice of compliance in this Court within
twenty-one days after issuance of this order. Defendant Lee shall make
and maintain written records of all actions taken pursuant to this Order
sufficient to document compliance with all requirements of this Order.
4. Pursuant to this Court’s authority to monitor the status of its injunction, when
the State of Florida adopts a new permanent ballot order scheme to replace
the ballot order scheme described in section 101.151(3)(a), Florida Statutes,
Defendant Lee shall file a notice in this Court informing this Court that the
State of Florida has done so, and shall append thereto a copy of the final text
of the ballot order scheme adopted. If the State of Florida, through Defendant
Lee or by any other means, adopts a temporary or interim ballot order scheme
to replace the ballot order scheme described in section 101.151(3)(a),
Defendant Lee shall file a notice in this Court informing this Court the State
of Florida has done so, and shall append thereto a copy of the temporary or
interim provision adopted.
73
Case 4:18-cv-00262-MW-CAS Document 202 Filed 11/15/19 Page 74 of 74
5. The Clerk shall close the file.
SO ORDERED on November 15, 2019.
s/Mark E. Walker
____
Chief United States District Judge
74
Disclaimer: Justia Dockets & Filings provides public litigation records from the federal appellate and district courts. These filings and docket sheets should not be considered findings of fact or liability, nor do they necessarily reflect the view of Justia.
Why Is My Information Online?